FC5 not properly configuring modem ofmultifunctionmodem/ethernet card

Komuro komurojun-mbn at nifty.com
Fri Jul 14 08:29:53 EDT 2006


Hello, 
 
> Yes, I've seen that. I think my case is different; the problem is that
> no driver is never loaded for the second card function, so I humbly
> deduce that the problem is with pcmciautils.

I think this is a bug of the kernel.

Please add the line "no_funcs = 2" to linux/drivers/pcmcia/ds.c
like below.
Actually, this is not a patch, but a test code.
(my kernel version 2.6.17)

without this test code 
>eth0: 3Com 3c589, io 0x300, irq 3, hw_addr 00:00:86:xx:xxl:xx
>  8K FIFO split 5:3 Rx:Tx, auto xcvr
>eth0: flipped to 10baseT

with this test code
>1.1: ttyS1 at I/O 0x2f8 (irq = 3) is a 16550A
>eth0: 3Com 3c589, io 0x300, irq 3, hw_addr 00:00:86:xx:xx:xx
>  8K FIFO split 5:3 Rx:Tx, auto xcvr
>eth0: flipped to 10baseT


static int pcmcia_card_add(struct pcmcia_socket *s)
{
	cisinfo_t cisinfo;
	cistpl_longlink_mfc_t mfc;
	unsigned int no_funcs, i;
	int ret = 0;

	if (!(s->resource_setup_done))
		return -EAGAIN; /* try again, but later... */

	if (pcmcia_validate_mem(s))
		return -EAGAIN; /* try again, but later... */

	ret = pccard_validate_cis(s, BIND_FN_ALL, &cisinfo);
	if (ret || !cisinfo.Chains) {
		ds_dbg(0, "invalid CIS or invalid resources\n");
		return -ENODEV;
	}

	if (!pccard_read_tuple(s, BIND_FN_ALL, CISTPL_LONGLINK_MFC, &mfc))
		no_funcs = mfc.nfn;
	else
		no_funcs = 1;

+	printk("no_funcs = %x \n", no_funcs);
+
+	no_funcs = 2;

	for (i=0; i < no_funcs; i++)
		pcmcia_device_add(s, i);

	return (ret);
}

Best Regards
Komuro




More information about the linux-pcmcia mailing list