Bug #556

usb stick/hub detach panic

Added by corecode almost 8 years ago. Updated 6 months ago.

Status:FeedbackStart date:
Priority:NormalDue date:
Assignee:tuxillo% Done:

0%

Category:Driver
Target version:3.8.0

Description

hey,

100% reproducible panic on detach of my usb stick. ehci is loaded as well. UP system.

#12 0xc0194734 in kprintf (fmt=0xdeadc0de <Address 0xdeadc0de out of bounds>)
at /usr/build/src/sys/kern/subr_prf.c:304
#13 0xc0260cfd in usb_disconnect_port (up=0xc415a178, parent=0xdeadc0de)
at /usr/build/src/sys/bus/usb/usb_subr.c:1382
#14 0xc02621a3 in uhub_detach (self=0xc415a100) at /usr/build/src/sys/bus/usb/uhub.c:610
#15 0xc018d9ae in device_detach (dev=0xc415a100) at device_if.h:48
#16 0xc018cef9 in device_delete_child (dev=0xdeadc0de, child=0xc415a100)
at /usr/build/src/sys/kern/subr_bus.c:591
#17 0xc0260d4b in usb_disconnect_port (up=0xc698b1f4, parent=0xdeadc0de)
at /usr/build/src/sys/bus/usb/usb_subr.c:1387
#18 0xc0261f33 in uhub_explore (dev=0xc4139940) at /usr/build/src/sys/bus/usb/uhub.c:460
#19 0xc025c6b7 in usb_discover (v=0xc695d9e0) at /usr/build/src/sys/bus/usb/usb.c:780
#20 0xc025bf71 in usb_event_thread (arg=0xc695d9e0) at /usr/build/src/sys/bus/usb/usb.c:469

gdb is broken somehow, the 0xdeadc0de arguments are not true, except for for kprintf.

(kgdb) fra 13
#13 0xc0260cfd in usb_disconnect_port (up=0xc415a178, parent=0xdeadc0de)
at /usr/build/src/sys/bus/usb/usb_subr.c:1382
1382 kprintf("%s: at %s", USBDEVPTRNAME(dev->subdevs[i]),
(kgdb) p dev->subdevs[0]
$1 = 0xc415a190
(kgdb) p *dev->subdevs[0]
$2 = {ops = 0xc415abf8, link = {tqe_next = 0xdeadc0de, tqe_prev = 0xdeadc0de},
parent = 0xdeadc0de, children = {tqh_first = 0xdeadc0de, tqh_last = 0xdeadc0de},
driver = 0xdeadc0de, devclass = 0xdeadc0de, unit = -559038242,
nameunit = 0xdeadc0de <Address 0xdeadc0de out of bounds>,
desc = 0xdeadc0de <Address 0xdeadc0de out of bounds>, busy = -559038242, state = 3735929054,
devflags = 3735929054, flags = 49374, order = 173 '�', pad = 222 '�, ivars = 0xdeadc0de,
softc = 0x0}

this stick has an integrated usb hub.

dump arriving on leaf at ~corecode/crash/*13*

cheers
simon

History

#1 Updated by sepherosa almost 8 years ago

No further child deletion means uhub has no children (i.e. umass) at
this point. Children probably detached long before we reach here but
uhub was not informed about the detaching.

Please
1) uncomment uhub.c line 744 recompile kernel and modules.
uhub_child_detached() is the place where uhub is informed about
children detaching. If child is not found in the upper loop, then
things probably go very wrong :(
2) give me the dmesg after you plugin the stick
3) give me the message printed on console immediately before panic

Best Regards,
sephe

#2 Updated by alexh over 5 years ago

Does this bug still occur? Did anyone experience it lately? Is it worth keeping
this bug report open?

#3 Updated by corecode over 5 years ago

Alex Hornung (via DragonFly issue tracker) wrote:
> Does this bug still occur? Did anyone experience it lately? Is it worth keeping
> this bug report open?

I lost the USB stick, but I'm sure the bug still exists.

cheers
simon

#4 Updated by dillon over 5 years ago

:
:Alex Hornung (via DragonFly issue tracker) wrote:
:> Does this bug still occur? Did anyone experience it lately? Is it worth=
: keeping=20
:> this bug report open?
:
:I lost the USB stick, but I'm sure the bug still exists.
:
:cheers
: simon

I still get panics on boot when my USB keyboard is attached to certain
ports. It could be related.

-Matt

#5 Updated by tuxillo 9 months ago

  • Description updated (diff)
  • Category set to Driver
  • Assignee changed from 0 to tuxillo
  • Target version set to 3.8.0

Grab.

#6 Updated by tuxillo 6 months ago

  • Status changed from New to Feedback

Hi all,

Is there any way we can check this at all? We're using U4B now.
Matt, do you still have that keyboard/computer?

Best regards,
Antonio Huete

Also available in: Atom PDF