Re: [sbp2] external hard disk not detected

[Stefan R. <st...@s5...>]

Fabrice Eudes wrote on 2006-04-24:
> I have a "Gericom Webgine 1500" under Debian GNU/Linux unstable. This
> hardware has a integrated Firewire 4-pin port.
> 
> I recently bought an external 3,5" hard disk enclosure (Advance
> BX362-COM) with two Firewire 6-pin ports in which i set a 200 GB Maxtor
> DiamondMax 10 hard disk. The enclosure also a USB 2.0 port but my laptop
> only has USB 1.1 ports, so i really want Firewire support as to achieve
> good data transfer rates.

There was a report quite some time ago about a USB/FireWire enclosure 
which switched from USB to FireWire only if ther was bus power present 
on the FireWire bus. All 4-pin ports and even some 6- and 9-pin ports 
(notably those of CardBus adapters) are unpowered. However since 
basically all Wintel notebooks have unpowered ports, this autoswitch 
hardware bug should be really obvious to hardware designers and 
therefore rare.

> I tried to follow the instructions on the http://www.linux1394.org/
> site, but without success :-(
> 
> My kernel is 2.6.15, compiled from the debian modified kernel sources. I
> do not know where to look to give the ieee1394 and the libraw1394 version :-/

The libraw version doesn't matter for device recognition by the kernel. 
The 1394 kernel driver version is probably the same as in Linus' kernel 
2.6.15 or perhaps of a 2.6.15.x stable release. (I don't know how Debian 
manages it.)

> I use the OHCI driver, and it looks fine to me: when I load the module,
> I get the following in /var/log/syslog:
> === start ===
> Apr 24 18:12:31 cassandre kernel: ohci1394: $Rev: 1313 $ Ben Collins <bco...@de...>
> Apr 24 18:12:31 cassandre kernel: ACPI: PCI Interrupt 0000:00:10.0[A] -> Link [LNKC] -> GSI 10 (level, low) -> IRQ 10
> Apr 24 18:12:31 cassandre kernel: ohci1394: fw-host0: OHCI-1394 1.0 (PCI): IRQ=[10]  MMIO=[dffff800-dfffffff]  Max Packet=[2048]
> Apr 24 18:12:32 cassandre kernel: ieee1394: Host added: ID:BUS[0-00:1023]  GUID[00030d4334125616]
> ==== end ====
> 
> The 'lspci' command and http://www.pcidatabase.com tell me that the
> controller is in fact a via one:
> 
> cassandre:~# lspci |grep FireWire
> 0000:00:10.0 FireWire (IEEE 1394): VIA Technologies, Inc. IEEE 1394 Host Controller (rev 46)
> cassandre:~# lspci -n |grep 10.0
> 0000:00:10.0 0c00: 1106:3044 (rev 46)
> 
> "0x1106","0x3044","VIA Technologies Inc","VT63061","VIA Fire II 1394a OHCI Link Layer Ctrlr1"
> 
> Also, the controller looks fine in gscanbus after loading the raw1394
> module.

Do you have the disk already attached at this point? If yes, gscanbus 
should show the disk in addition to the controller.

> The problem is that the hard disk doesn't seem to be detected when i
> load the sbp2 module; in /var/log/syslog, I get:
> 
> === start ===
> Apr 24 18:24:02 cassandre kernel: SCSI subsystem initialized
> Apr 24 18:24:02 cassandre kernel: sbp2: $Rev: 1306 $ Ben Collins <bco...@de...>
> Apr 24 18:24:02 cassandre kernel: ieee1394: sbp2: Driver forced to serialize I/O (serialize_io=1)
> Apr 24 18:24:02 cassandre kernel: ieee1394: sbp2: Try serialize_io=0 for better performance
> ==== end ====
> 
> 1) Can you confirm me that it is not a udev problem: the disk should
> appear in syslog whether i use hotplug or udev doesn't matter ?

Although I don't know if udev is working properly, I can confirm that 
the problem lies much deeper. The ieee1394 base driver should log a line 
like
kernel: ieee1394: Node added: ID:BUS[0-00:1023]  GUID[0001d202e08500e7]
as soon as the disk is plugged in. After that, a userspace helper like 
hotplug or hal should load the sbp2 driver.

But you log not only lacks this log line but even any error message that 
could appear instead.

> 2) Any advice (or complete solution ;-) ) to my problem ?

Does gscanbus show anything?
-- 
Stefan Richter
-=====-=-==- -=-- ==-=-
http://arcgraph.de/sr/