Bug 969

Summary:	"rmmod uhci-hcd" - thread is blocked forever in kernel
Product:	Drivers	Reporter:	Attila Kesmarki (danthe)
Component:	USB	Assignee:	Greg Kroah-Hartman (greg)
Status:	REJECTED DUPLICATE
Severity:	normal
Priority:	P2
Hardware:	i386
OS:	Linux
Kernel Version:	2.6.0-test1	Subsystem:
Regression:	---	Bisected commit-id:
Attachments:	boot messages dmesg output after blocking rmmod dmesg output after rmmod, lsmod, modprobe (3 blocked tasks) my kernel .config file the loaded modules right before executing rmmod uhci-hcd

Description Attila Kesmarki 2003-07-20 05:32:40 UTC

Distribution: Gentoo  
Hardware Environment: Athlon xp2000 + Gigabyte GA-7VAXP Motherboard 
Software Environment: 
2.6.0-test1 kernel with ACPI & USB compiled in as modules connected to a Compaq 
ipaq h3950 Pocketpc 
 
Problem Description: 
With my USB device attached, the linux can't shutdown itself properly. 
I think I have found it's cause: a thread in the kernel is blocked, and it doesn't react 
to SIGTERM/SIGKILL.   
I got this (I think) problem without rebooting, when I wanted to "rmmod uhci-hcd": 
   
danhome root # sync 
danhome root # rmmod uhci_hcd 
uhci-hcd 0000:00:10.0: remove, state 3 
usb usb2: USB disconnect, address 1 
usb 2-1: USB disconnect, address 2 
 ---------- The execution blocks here forever, I don't get the prompt back --------- 
 
After this, I've dumped  the call stacks of each tasks, I will attach this file. 
After executing this command, I can't kill the blocked rmmod task, furthermore the 
lsmod, modprobe, and the cat /proc/modules will be blocked, as well. (locking 
problems caused by the rmmod task, I think) Anyway, I will attach an other  file with 
those processes' call stacks. 
 
Steps to reproduce: 
If I just plug in my (active, switched on) pocketPC into the USB port of my PC, boot 
the Linux, and try to rmmod uhci-hcd, the problem is reproduced.

Comment 1 Attila Kesmarki 2003-07-20 05:34:11 UTC

Created attachment 583 [details]
boot messages

Comment 2 Attila Kesmarki 2003-07-20 05:34:45 UTC

Created attachment 584 [details]
dmesg output after blocking rmmod

Comment 3 Attila Kesmarki 2003-07-20 05:35:23 UTC

Created attachment 585 [details]
dmesg output after rmmod, lsmod, modprobe (3 blocked tasks)

Comment 4 Attila Kesmarki 2003-07-20 05:36:32 UTC

Created attachment 586 [details]
my kernel .config file

Comment 5 Attila Kesmarki 2003-07-20 05:37:26 UTC

Created attachment 587 [details]
the loaded modules right before executing rmmod uhci-hcd

Comment 6 Attila Kesmarki 2003-07-20 05:45:29 UTC

Additional info: 
 
------------------------------------------------------------------------- 
 
danhome root # scanpci 
 
pci bus 0x0000 cardnum 0x00 function 0x00: vendor 0x1106 device 0x3189 
 VIA Technologies, Inc. VT8377 [KT400 AGP] Host Bridge 
 
pci bus 0x0000 cardnum 0x01 function 0x00: vendor 0x1106 device 0xb168 
 VIA Technologies, Inc. VT8235 PCI Bridge 
 
pci bus 0x0000 cardnum 0x09 function 0x00: vendor 0x1113 device 0x1216 
 Accton Technology Corporation EN-1216 Ethernet Adapter 
 
pci bus 0x0000 cardnum 0x10 function 0x00: vendor 0x1106 device 0x3038 
 VIA Technologies, Inc. USB 
 
pci bus 0x0000 cardnum 0x10 function 0x01: vendor 0x1106 device 0x3038 
 VIA Technologies, Inc. USB 
 
pci bus 0x0000 cardnum 0x10 function 0x02: vendor 0x1106 device 0x3038 
 VIA Technologies, Inc. USB 
 
pci bus 0x0000 cardnum 0x10 function 0x03: vendor 0x1106 device 0x3104 
 VIA Technologies, Inc. USB 2.0 
 
pci bus 0x0000 cardnum 0x11 function 0x00: vendor 0x1106 device 0x3177 
 VIA Technologies, Inc. VT8235 ISA Bridge 
 
pci bus 0x0000 cardnum 0x11 function 0x01: vendor 0x1106 device 0x0571 
 VIA Technologies, Inc. VT82C586/B/686A/B PIPC Bus Master IDE 
 
pci bus 0x0000 cardnum 0x11 function 0x05: vendor 0x1106 device 0x3059 
 VIA Technologies, Inc. VT8233 AC97 Audio Controller 
 
pci bus 0x0000 cardnum 0x13 function 0x00: vendor 0x10ec device 0x8139 
 Realtek Semiconductor Co., Ltd. RTL-8139/8139C/8139C+ 
 
pci bus 0x0001 cardnum 0x00 function 0x00: vendor 0x1002 device 0x4966 
 ATI Technologies Inc Radeon R250 If [Radeon 9000] 
 
--------------------------------------------------------------------- 
 
danhome root # cat /proc/interrupts 
           CPU0 
  0:    2838794    IO-APIC-edge  timer 
  1:      12714    IO-APIC-edge  i8042 
  2:          0          XT-PIC  cascade 
  8:          2    IO-APIC-edge  rtc 
  9:          0   IO-APIC-level  acpi 
 10:          0    IO-APIC-edge  ehci_hcd, uhci-hcd 
 11:          0    IO-APIC-edge  VIA8233, uhci-hcd, uhci-hcd 
 12:     118250    IO-APIC-edge  i8042 
 14:      25162    IO-APIC-edge  ide0 
 15:         19    IO-APIC-edge  ide1 
 16:     197105   IO-APIC-level  radeon@PCI:1:0:0 
 17:       4678   IO-APIC-level  eth1 
 18:         38   IO-APIC-level  eth0 
NMI:          0 
LOC:    2838838 
ERR:          0 
MIS:          0

Comment 7 Greg Kroah-Hartman 2003-07-21 10:07:22 UTC

What does the kernel log messages show when you run rmmod?

Do you have any usb devices connected to your system?  What kind?

Comment 8 Attila Kesmarki 2003-07-21 18:35:27 UTC

 
Okay, I've found the cause of this bug. When I disabled ACPI using "noacpi 
pci=noacpi" as kernel parameters, the bug disappeared. 
As my motherboard is a GA-7VAXP, I'm 99% sure that this is the same as bug#845 
( which is the  umbrella bug of bug#10, #71, and #678 )

Comment 9 Greg Kroah-Hartman 2003-07-22 06:06:21 UTC


*** This bug has been marked as a duplicate of 10 ***