[hsflinux] my modem just segmentations fault when I attempt to open the device...

[Jeff Wiegley, Ph.D.]

On Wed, 2002-05-29 at 22:48, Tony Earnshaw wrote:
> tor, 2002-05-30 kl. 00:31 skrev Jeff Wiegley, Ph.D.:
> 
> > I've got a Dell Inspiron 4100 laptop and it has an Actiontec
> > MD56ORD internal modem.  I think its a Conexant HSF based
> > device and I'ld love to get it working. (So would all the
> > other Inspiron owners I'm sure).
> 
> I have a different notebook, Compaq Presario 700EA with RH 7.2, kernel
> 2.4.18. HSFi modem, everything works - otherwise you wouldn't be reading
> this.
> 
> Offhand I can think of 3 possibilities for your problem, but I'm not the
> expert here.
> 
> >	Subsystem: Conexant: Unknown device 5421
> > 	Flags: bus master, medium devsel, latency 0, IRQ 11
> > 	I/O ports at d400 [size=256]
> > 	I/O ports at dc00 [size=128]
> 
> 1: IRQ 11 isn't normal for a modem, but can be negotiated automatically
> when the modules are loaded. Modems should use COM port / tty IRQs,
> generally 4 and 3. My machine suggests IRQ 11 for card bus and Ethernet.
> 
> > [root at localhost root]# ls -al /dev/ttySHSF0 
> > crw-------    1 root     root     240,  64 Dec 31  1969 /dev/ttySHSF0
> 
> 2: The permissions are wrong. Don't know why they should be, hsfconfig
> should have set them to 644. You can safely do this by hand, as root.

I'll take a look at the log files and see if the IRQ is being
detected/set properly. The name of the tty doesn't matter only the
major/minor number

I don't think the permissions should matter because all the testing I'm
doing is being done as root so 600 and 644 should be effectively the
same.

The log files show this:
May 30 09:27:06 localhost insmod: Warning: loading
/lib/modules/2.4.19-pre8/misc/hsfich.o will taint the kernel: non-GPL
license - Copyright (C) 1996-2001 Conexant Systems Inc. All Rights
Reserved.
May 30 09:27:06 localhost insmod:   See http://www.tux.org/lkml/#s1-18
for information about tainted modules
May 30 09:27:06 localhost insmod: Module hsfich loaded, with warnings
May 30 09:27:06 localhost kernel: PCI: Found IRQ 11 for device 00:1f.6
May 30 09:27:06 localhost kernel: PCI: Sharing IRQ 11 with 00:1f.5
May 30 09:27:06 localhost kernel: PCI: Sharing IRQ 11 with 02:03.0
May 30 09:27:06 localhost kernel: PCI: Setting latency timer of device
00:1f.6 to 64

So it did pick up the right IRQ. But that IRQ is being shared with
two other devices:

00:1f.5 Multimedia audio controller: Intel Corp. 82801CA/CAM AC'97 Audio
(rev 01)
	Subsystem: Cirrus Logic: Unknown device 5959
	Flags: bus master, medium devsel, latency 0, IRQ 11
	I/O ports at d800 [size=256]
	I/O ports at dc80 [size=64]

02:03.0 CardBus bridge: Texas Instruments PCI1410 PC card Cardbus
Controller (rev 01)
	Subsystem: Lucent Technologies: Unknown device ab01
	Flags: bus master, medium devsel, latency 168, IRQ 11
	Memory at f4002000 (32-bit, non-prefetchable) [size=4K]
	Bus: primary=02, secondary=0b, subordinate=0e, sec-latency=176
	Memory window 0: f5400000-f57ff000 (prefetchable)
	Memory window 1: f5800000-f5bff000
	I/O window 0: 0000f400-0000f4ff
	I/O window 1: 0000f800-0000f8ff
	16-bit legacy interface ports at 0001


> 
> 3: I don't like pre-release kernels. Is there any special reason why you
> are running 2.4.19? For example, latest iptables patches demand this,
> but if you don't have to run 2.4.19, get the 2.4.18 source from
> kernel.org (*NOT* Red Hat) and compile and install that (with its own
> kernel modules, and compile the HSF modules against its source code and
> libraries). You can safely boot on either kernel if you modify lilo.conf
> or the grub thing properly, so you're not sacrificing anything. I doubt
> whether Marc has used 2.4.19 for any of his drivers.
> 

This exhibits the same behavior on other non-prerelease kernels as well.

> Lastly, 2 things:
> 
> - Your Dell doesn't seem to support ACPI power control, which is a pity;

My Dell does support it but my kernel does not. My knowledge of ACPI is
that it isn't as functional as APM. Plus I don't know how to configure
ACPI stuff and I haven't been able to find a good howto on the subject.
So I disable ACPI in the kernel and use APM instead. If you know of a
good howto or can point in the right direction to learn how to properly
use ACPI and configure as much ACPI functionality as I can I would be
willing to spend the time to switch.  What are the advantages to ACPI?

[We should carry this discussion on external to the hsf mailing list
though].

> - I've compiled my different kernels with both gcc 2.96 (Red Hat
> bastardization) and 3.04 and both work fine. I've also compiled the HSF
> modules with 2.96 and 3.04, and both work fine. However, the latest
> GNUPG will not compile properly with gcc 3.04, so your gcc version just
> could have some influence on the whole.

I've just been sticking with 2.96 since it ships with redhat 7.3  and
gcc3 doesn't. There are enough items out that don't compile under gcc3
that I don't want to waste my time having to figure out that its the
compiler that is preventing things from working.

> > Hope this helps,

The IRQ item might help I'll check that out later today.

> 
> Tony
> 
> -- 
> 
> Tony Earnshaw
> 
> e-post:		tonni at billy.demon.nl
> www:		http://www.billy.demon.nl
> gpg public key:	http://www.billy.demon.nl/tonni.armor
> 
> Telefoon:	(+31) (0)172 530428
> Mobiel:		(+31) (0)6 51153356
> 
> GPG Fingerprint = 3924 6BF8 A755 DE1A 4AD6 FA2B F7D7 6051 3BE7 B981
> 3BE7B981
> 
> 

- Jeff