Bug 100533

Summary: USB memory stick not recognized - modprobe helps
Product: [openSUSE] SUSE LINUX 10.0 Reporter: J. Daniel Schmidt <jdsn>
Component: KernelAssignee: Greg Kroah-Hartman <gregkh>
Status: RESOLVED DUPLICATE QA Contact: E-mail List <qa-bugs>
Severity: Normal    
Priority: P5 - None    
Version: Beta 2   
Target Milestone: ---   
Hardware: i386   
OS: All   
Whiteboard:
Found By: Development Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---
Attachments: Potential patch for this issue.

Description J. Daniel Schmidt 2005-08-03 15:26:52 UTC 
I just plugged a TrekStor USB 2.0 memory stick into my workstation.
As it was not recognized automatically I fetched the following from
/var/log/messages

---- snip ----
Aug  3 16:20:55 linux kernel: usb 5-1: new high speed USB device using ehci_hcd
and address 9
Aug  3 16:20:55 linux kernel: usb 5-1: device descriptor read/64, error -71
Aug  3 16:20:55 linux kernel: usb 5-1: device descriptor read/64, error -71
Aug  3 16:20:55 linux kernel: usb 5-1: new high speed USB device using
ehci_hcd and address 10
Aug  3 16:20:55 linux kernel: usb 5-1: device descriptor read/64, error -71
Aug  3 16:20:55 linux kernel: usb 5-1: device descriptor read/64, error -71
Aug  3 16:20:56 linux kernel: usb 5-1: new high speed USB device using
ehci_hcd and address 11
Aug  3 16:20:56 linux kernel: usb 5-1: device not accepting address 11,
error -32
Aug  3 16:20:56 linux kernel: usb 5-1: new high speed USB device using
ehci_hcd and address 12
Aug  3 16:20:56 linux kernel: usb 5-1: device descriptor read/8, error -71
Aug  3 16:20:56 linux kernel: usb 5-1: device descriptor read/8, error -71
---- snip ----


After a 
  # modprobe -r ehci-hcd
and plugging in the device again, everything worked fine.


This behaviour was detected on two identical machines (belana and e228), both
Intel P4. The stick worked without any problems on a AMD64 machine (tested on
pdbdevel).


According to http://www.ubuntu-forum.de/thread.php?threadid=1077 this seems to
be a kernel problem (on that page search for "sudo modprobe -r ehci-hcd").

I hope this helps to solve the problem. (I did not test it with a preview of 10.0).
Comment 1 Olaf Hering 2005-08-05 14:55:35 UTC 
I suggest to try 10.0 preview4, just boot the cd and load usb-storage.
using the bootiso and boot with 'insmod=usb-storage' should be enough.
Comment 2 J. Daniel Schmidt 2005-08-09 13:01:57 UTC 
Problem still persists with beta1 (aka Prague).
Comment 3 Greg Kroah-Hartman 2005-08-10 18:34:02 UTC 
If you unload the ehci-hcd driver, does this fix the problem (yes, it will
go slower, but trying to determine if this is an interrupt routing problem or not.)
Comment 4 J. Daniel Schmidt 2005-08-11 15:34:48 UTC 
Yea, if I unload the driver first and plug in the stick after that it works fine.
Comment 5 Greg Kroah-Hartman 2005-08-12 00:23:15 UTC 
Ok, if you change
/sys/module/usbcore/parameters/old_scheme_first to "Y" by doing:
  echo 1 > /sys/module/usbcore/parameters/old_scheme_first 

and then trying with the ehci-hcd driver loaded, does that work?
Comment 6 J. Daniel Schmidt 2005-08-12 12:02:32 UTC 
No, with that change it does not work!


But something else has changed :)
If I do now a 
   # modprobe -r ehci-hcd
the stick is recognized right after that and I don't have to plug it out and in
again (as before).
Comment 7 Greg Kroah-Hartman 2005-08-16 08:01:23 UTC 
Yeah, but that means it is slower than need be, which is not a good thing.

Did the error messages change with old_scheme_first changed to Y?
Comment 8 J. Daniel Schmidt 2005-08-16 11:21:16 UTC 
No, the messages did not change. I meant that by saying: 
"No, with that change it does not work!"
Comment 9 J. Daniel Schmidt 2005-08-17 18:22:51 UTC 
Problem still persists with beta2 (aka Boston).
Comment 10 Greg Kroah-Hartman 2005-08-18 04:05:40 UTC 
Created attachment 46389 [details]
Potential patch for this issue.

Can you test this patch out to see if it solves the problem?
Comment 11 Olaf Hering 2005-08-20 11:42:26 UTC 
try these rpms:

ftp.suse.com/pub/people/olh/kernel/bug100533

201926267104fe748e3b0d1d326dd031  kernel-default-2.6.13_rc6_git11-3.i586.rpm
860d161201fba2c4fad6b48cf624eaf7  kernel-default-2.6.13_rc6_git11-3.nosrc.rpm
94d51c88cb6824b7c530eed7c2d7b196  kernel-default-2.6.13_rc6_git11-3.x86_64.rpm
d6dfd797324528432bfd9161ee16ae62  kernel-default-nongpl-2.6.13_rc6_git11-3.i586.rpm
03205794cbc333aee33f4f35802ffb94 
kernel-default-nongpl-2.6.13_rc6_git11-3.x86_64.rpm
a67a3a815015b3ac32fbb52df3b6fb38  kernel-smp-2.6.13_rc6_git11-3.i586.rpm
abcef07b45790a1ef8f3cf9277914dcd  kernel-smp-2.6.13_rc6_git11-3.nosrc.rpm
7f7acb53f5d48e771b76ae942e664cd7  kernel-smp-2.6.13_rc6_git11-3.x86_64.rpm
93f56a525ea90b0aa9c4c5e6a3199c88  kernel-smp-nongpl-2.6.13_rc6_git11-3.i586.rpm
abf6baec6724f8d1ab908c95837c2b11  kernel-smp-nongpl-2.6.13_rc6_git11-3.x86_64.rpm
Comment 12 J. Daniel Schmidt 2005-08-22 15:57:50 UTC 
Thanks for preparing the kernel rpms.

I am sorry to say, that this kernel patch changed nothing. The same error
messages appear in /var/log/messages .
Comment 13 J. Daniel Schmidt 2005-09-06 14:19:09 UTC 
Seems to be the same problem but with different hardware
According to ro@suse.de this is a problem with the USB 2.0 chipset drivers ...

*** This bug has been marked as a duplicate of 105964 ***