Bug 411241

Summary:	USB drive doesn't work
Product:	Red Hat Enterprise Linux 4	Reporter:	starlight
Component:	kernel	Assignee:	Pete Zaitcev <zaitcev>
Status:	CLOSED WORKSFORME	QA Contact:	Martin Jenner <mjenner>
Severity:	high	Docs Contact:
Priority:	low
Version:	4.5
Target Milestone:	---
Target Release:	---
Hardware:	x86_64
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2008-01-15 01:04:14 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description starlight 2007-12-04 23:03:53 UTC

Description of problem:

This broke sometime since RHEL 4U1, or possibly
due to a change of mainboard to Tyan S2866.

Upon inserting USB connection 'dmesg' shows

usb 1-1: new high speed USB device using address 2
usb 1-1: control timeout on ep0in
usb 1-1: device not accepting address 2, error -71

And drive fails to insert.  Googled the above and
found that running

rmmod ehci_hcd

first fixes the problem.

Version-Release number of selected component (if applicable):

2.6.9-55.0.12.ELsmp

Comment 1 starlight 2007-12-05 00:48:59 UTC

Forget about the strange 'rmmod' workaround, the USB disk 
drivers do not work anymore.  Tried copying a bunch of big files 
to the disk.  Data write rate per 'iostat' was a paltry 1MB/sec 
until 1.5GB of data had been copied.  Then it screeched to a 
dead stop.  'iostat' said that 0 kb/s were being transferred at 
9 tps.  This went on for 30 minutes.  A 'sync' command just 
hung. Had to unplug the USB cable to get it out of the bizarre 
state.

This definitely used to work!

Gave up on RHEL and plugged the drive into a Windows 2003 R2-SP2 
X64 server.  Formatted the drive as NTFS and the copy is now 
humming along nicely at 16.5MB/sec.

RHEL is a commercial grade OS?  Not in my book.  This is an 
amateur regression if ever there was one.  Also just finished
wrestling with 'st' tape driver regressions, bug 410231 and
bug 411131.  Might be time to try out SLES.  I know a couple
of companies that have made the switch.

Comment 2 Pete Zaitcev 2008-01-15 01:04:14 UTC

Looks like solution was found, closing.

Comment 3 starlight 2008-01-15 01:49:42 UTC

Not a Red Hat solution however.