Bug 6708 - ehci_hcd hangs on insertion of usb device
ehci_hcd hangs on insertion of usb device
Status: REJECTED INSUFFICIENT_DATA
Product: Drivers
Classification: Unclassified
Component: USB
i386 Linux
: P2 normal
Assigned To: David Brownell
:
Depends on:
Blocks: USB
  Show dependency treegraph
 
Reported: 2006-06-18 05:09 UTC by dev
Modified: 2007-09-23 17:15 UTC (History)
3 users (show)

See Also:
Kernel Version: 2.6.16
Tree: Mainline
Regression: Yes


Attachments
/proc/bus/usb/devices (2.83 KB, text/plain)
2006-06-20 02:28 UTC, dev
Details
output of lspci -v (4.63 KB, text/plain)
2006-06-20 02:32 UTC, dev
Details
dmesg output from the 1.) case (20.42 KB, text/plain)
2006-09-10 17:58 UTC, dev
Details
dmesg output from the 2.) case (12.13 KB, text/plain)
2006-09-10 17:59 UTC, dev
Details
lsusb -v (uhci only) (12.87 KB, text/plain)
2006-09-10 18:00 UTC, dev
Details

Description dev 2006-06-18 05:09:11 UTC
Most recent kernel where this bug did not occur:2.6.15 series
Distribution:Gentoo Linux
Hardware Environment:x86, athlon-xp, via-kt400
Software Environment:
Problem Description:
The troublemaker is an Trekstor iBeat Organix device. It used to work just fine
under the 2.6.15 kernel series but since 2.6.16 the ohci_hcd driver hangs on
insertion of the device. rmmod ohci_hcd and the device works just fine with
uhci_hcd. The funny thing is that the whole usb subsystem becomes unusable on
insertion of the device. lsusb hangs, tail /proc/usb/devices hangs, removing or
inserting any other usb device during the hang does not work either, ...
After removal of the Organix everything starts working just fine again.

dmesg output:

hub 1-0:1.0: state 7 ports 6 chg 0000 evt 0002
ehci_hcd 0000:00:10.3: GetStatus port 1 status 001803 POWER sig=j CSC CONNECT
hub 1-0:1.0: port 1, status 0501, change 0001, 480 Mb/s
hub 1-0:1.0: debounce: port 1: total 100ms stable 100ms status 0x501
ehci_hcd 0000:00:10.3: port 1 high speed
ehci_hcd 0000:00:10.3: GetStatus port 1 status 001005 POWER sig=se0 PE CONNECT
usb 1-1: new high speed USB device using ehci_hcd and address 5
ehci_hcd 0000:00:10.3: port 1 high speed
ehci_hcd 0000:00:10.3: GetStatus port 1 status 001005 POWER sig=se0 PE CONNECT
usb 1-1: default language 0x0409
usb 1-1: new device strings: Mfr=1, Product=2, SerialNumber=3
usb 1-1: Product: i.Beat organix
usb 1-1: Manufacturer: TrekStor
usb 1-1: SerialNumber: 0002F68172B2CC9D
usb 1-1: uevent
usb 1-1: device is bus-powered
usb 1-1: configuration #1 chosen from 1 choice
usb 1-1: hcd_unlink_urb f6de3860 fail -16


Steps to reproduce:
modprobe ohci_hcd
insert iBeat Organix device
Comment 1 Andrew Morton 2006-06-19 19:47:36 UTC
On Mon, 19 Jun 2006 19:21:50 -0700
David Brownell <david-b@pacbell.net> wrote:

> On Monday 19 June 2006 3:47 pm, Andrew Morton wrote:
> > 
> > yaregression
> 
> Or at least a severe problem between brain and keyboard.
> 
> Get this:  a system with VIA kt-400 hardware,
> which ** DOES NOT INCLUDE OHCI ** is getting some
> problem ... which is then blamed on the OHCI driver.
> 
> And the "bug report" doesn't show anything having
> to do with OHCI either.
> 
> Even if this _were_ a problem with that driver,
> this is a completely unusable bug report.
> 

OK, but "rmmod ohci_hcd and the device works just fine with uhci_hcd" seems
pretty unambiguous.

So the reporter appears to be saying that the mere presence of that module
makes the machine fail, despite (or because of?) the absence of that
driver's hardware?

Comment 2 Anonymous Emailer 2006-06-19 20:50:01 UTC
Reply-To: david-b@pacbell.net

On Monday 19 June 2006 7:48 pm, Andrew Morton wrote:

> 
> So the reporter appears to be saying that the mere presence of that module
> makes the machine fail, despite (or because of?) the absence of that
> driver's hardware?

Which would mean some problem in PCI or the module loader, not USB.

And yet the bug report showed no evidence that the OHCI driver was
even loaded ... like, say, its banner message.  Hence the comment
that the bug report is broken.  There are FAQs on how to make a
bug report, including one focussed on USB; this one ignores them all.



Comment 3 dev 2006-06-20 02:25:24 UTC
Well not exactly brain failure but somewhat blind and typing from memory. To
clearify I meant the ehci_hcd driver making troubles. Please be so kind to
query-replace all occurences of ohci_hcd with ehci_hcd. As to the rest of the
missing information ill attach it shortly. If u need more information feel free
to ask.
Comment 4 dev 2006-06-20 02:28:46 UTC
Created attachment 8349 [details]
/proc/bus/usb/devices
Comment 5 dev 2006-06-20 02:32:22 UTC
Created attachment 8350 [details]
output of lspci -v
Comment 6 David Brownell 2006-09-10 11:26:28 UTC
does this still happen with 2.6.18-rc6? 
 
Missing information includes descriptors of the device 
that appears to be triggering the problem ('lsusb -v' 
is best) ... notice that the /proc/bus/usb/devices info 
doesn't show that device!! 
 
Missing information also includes dmesg traces, from the 
initializaiton of USB to after the problem appears, with 
CONFIG_USB_DEBUG enabled. 
 
Comment 7 dev 2006-09-10 17:54:52 UTC
Still hangs with 2.6.18_rc6 but sometimes succeeds. Though even if it succeeds 
running lsusb -v always hangs the device. Going to attach 2 dmesg outputs and 
lsusb -v output from only running with uhci module loaded.
dmesg outputs:
1.) a) Bootup sequence
    b) Connected device, succeeded
    c) Unplugged device
    d) Reconnected device, this time failed
2.) a) rmmoded ehci_hcd module & run dmesg -c
    b) modprobe ehci_hcd
    c) connected device, succeeded
    d) run lsusb -v, which hung
    e) disconnected device, lsusb -v unfroze at this point
Comment 8 dev 2006-09-10 17:58:09 UTC
Created attachment 8981 [details]
dmesg output from the 1.) case
Comment 9 dev 2006-09-10 17:59:50 UTC
Created attachment 8982 [details]
dmesg output from the 2.) case
Comment 10 dev 2006-09-10 18:00:41 UTC
Created attachment 8983 [details]
lsusb -v (uhci only)
Comment 11 Natalie Protasevich 2007-06-13 12:37:53 UTC
Good chance the problem has been resolved in newer kernels especially 2.6.22-rc4.
Can you please give it a try?
Thanks,
--Natalie
Comment 12 Adrian Bunk 2007-09-23 17:15:12 UTC
Please reopen this bug if it's still present with kernel 2.6.22.

Note You need to log in before you can comment on or make changes to this bug.