[libdv-dev] AV/C hangs under kernel 2.6

[Marko <mar...@hu...>]

Dan,

On Thu, Mar 10, 2005 at 08:46:39AM -0500, Dan Dennedy wrote:
> Right now, module removal of ieee1394 modules under kernel 2.6 is known
> to be problematic. None of the kernel messages shown are alarming
> except, of course, the Oops on card eject. None of the messages relate
> to AV/C deadlocking Kino. Can you test if transport control in dvcont or
> gscanbus works OK?

gscanbus became unkillable and entered a busy loop (presumably in the kernel)
either as the result of changing my camera from record mode (via power-off)
to play mode, or attempting to click the mouse on the camera icon.  During
this test, the following was displayed on stdout or stderr:

Error while reading from IEEE1394: : Resource temporarily unavailable
0/0x0000fffff000040c: read failed
Error while reading from IEEE1394: : Resource temporarily unavailable
0/0x0000fffff0000410: read failed
Error while reading from IEEE1394: : Resource temporarily unavailable
0/0x0000fffff0000414: read failed
Error while reading from IEEE1394: : Resource temporarily unavailable
1/0x0000fffff0000400: read failed
1/0x0000fffff0000400: wrong bus info block length
Error while reading from IEEE1394: : Resource temporarily unavailable
Error while reading from IEEE1394: : Resource temporarily unavailable
Could not read topologyMap
Error while reading from IEEE1394: : Resource temporarily unavailable
Error while reading from IEEE1394: : Resource temporarily unavailable
Could not read topologyMap
Error while reading from IEEE1394: : Resource temporarily unavailable
1/0x0000fffff0000400: read failed
1/0x0000fffff0000400: wrong bus info block length
Error while reading from IEEE1394: : Invalid argument
Could not read topologyMap
Error while reading from IEEE1394: : Resource temporarily unavailable
0/0x0000fffff000040c: read failed

After rebooting, "dvcont verbose play" displayed

node 0 type = 2

and hanged (but didn't consume any noticeable amount of processing power)
until I turned off the camera (it was in play mode).  After that point,
it printed:

node 0 AVC video recorder? no
node 0 AVC disk recorder? no
node 0 AVC tuner? no
node 0 AVC video camera? no
node 0 AVC video monitor? no
Could not find any AV/C devices on the 1394 bus.

Before the dvcont test, the relevant dmesg output was as follows:

ohci1394: $Rev: 1223 $ Ben Collins <bco...@de...>
PCI: Enabling device 0000:03:00.0 (0000 -> 0002)
ACPI: PCI interrupt 0000:03:00.0[A] -> GSI 11 (level, low) -> IRQ 11
PCI: Setting latency timer of device 0000:03:00.0 to 64
ohci1394: fw-host0: Unexpected PCI resource length of 1000!
uhci_hcd 0000:00:1d.1: suspend_hc
ohci1394: fw-host0: OHCI-1394 1.1 (PCI): IRQ=[11]  MMIO=[10800000-108007ff]  Max Packet=[2048]
uhci_hcd 0000:00:1d.2: suspend_hc
ieee1394: Host added: ID:BUS[0-00:1023]  GUID[00004c020000002d]

After turning on the camera, the following was logged:

ieee1394: Error parsing configrom for node 0-00:1023
ieee1394: Error parsing configrom for node 0-01:1023
ieee1394: The root node is not cycle master capable; selecting a new root node and resetting...
ieee1394: Node added: ID:BUS[0-00:1023]  GUID[0080880000f134e6]
ieee1394: Node changed: 0-00:1023 -> 0-01:1023
ieee1394: raw1394: /dev/raw1394 device initialized
atkbd.c: Keyboard on isa0060/serio0 reports too many keys pressed.

Nothing further was printed when dvcont was hanging.  After I turned off the
camera, the following was printed:

isa0060/serio0 reports too many keys pressed.
ohci1394: fw-host0: AT dma reset ctx=0, aborting transmission
ieee1394: Node changed: 0-01:1023 -> 0-00:1023
ieee1394: Node suspended: ID:BUS[0-00:1023]  GUID[0080880000f134e6]
atkbd.c: Keyboard on isa0060/serio0 reports too many keys pressed.

Do you need the lspci output, or is there anything else I could do to help
debug this, such as adding some debugging output to the 2.6.11.1 kernel?

This Intel Centrino based system has an Intel 82855PM chipset, Texas
Instruments PCI4520 Cardbus controller and NEC Corporation IEEE 1394 Host
Controller (rev 01).  AV/C has worked fine under 2.4 kernels using the
same hardware.

	Marko