[linux-dvb] flexcop-pci.ko insmod failure

[Patrick Boettcher]

Hi Alan,

On Tue, 28 Jun 2005, Alan Halverson wrote:
> New to the list and linux dvb.  It's been a little frustrating getting going, 
> but I now have a kernel with support for a pair of Air2PC 1st gen cards with 
> the bcm3510 frontend.

How does it work with the bcm3510?

> I'm using the 2.6.12-git9 kernel patch on Fedora Core 3.  Initially, the 
> flexcop-pci.ko module failed to load with this error:
>
> Jun 28 08:25:46 dualxeon kernel: kobject_register failed for Technisat/B2C2 
> FlexCop II/IIb/III PCI (-13)
> Jun 28 08:25:46 dualxeon kernel:  [<c021777d>] kobject_register+0x5d/0x70
> Jun 28 08:25:46 dualxeon kernel:  [<c02a3b82>] bus_add_driver+0x72/0xc0
> Jun 28 08:25:46 dualxeon kernel:  [<c02250d8>] pci_register_driver+0x98/0xb0
> Jun 28 08:25:46 dualxeon kernel:  [<c014bf72>] sys_init_module+0x132/0x1d0
> Jun 28 08:25:46 dualxeon kernel:  [<c0103a8f>] sysenter_past_esp+0x54/0x75
>
> Turns out that having a forward slash "/" in the driver name is illegal. 
> With the following diff, the flexcop-pci.ko modules loads correctly:
>
> --- linux-2.6.12-clean/drivers/media/dvb/b2c2/flexcop-pci.c     2005-06-17 
> 14:48:29.000000000 -0500
> +++ linux-2.6.12/drivers/media/dvb/b2c2/flexcop-pci.c   2005-06-28 
> 09:24:03.000000000 -0500
> @@ -357,7 +357,7 @@
> MODULE_DEVICE_TABLE(pci, flexcop_pci_tbl);
> static struct pci_driver flexcop_pci_driver = {
> -       .name = "Technisat/B2C2 FlexCop II/IIb/III PCI",
> +       .name = "Technisat B2C2 FlexCop II IIb III PCI",
>        .id_table = flexcop_pci_tbl,
>        .probe = flexcop_pci_probe,
>        .remove = flexcop_pci_remove,
>
> I'm sure flexcop-usb.c (and others) would need a similar diff also.

This must be a new rule (?). There are quite a few people (including me) 
that are using the driver with plain 2.6.12 (and previous versions) 
successfully.

Does someone know something about the syntax of pci_driver->name?

Anyway, I will change the name-field to the actual driver name in all my 
drivers as it seems the name-field of the *_driver-struct is used some 
places where such long names are inappropriate.

Thanks for pointing that out.

best regards,
Patrick.

--
   Mail: patrick.boettcher at desy.de
   WWW:  http://www.wi-bw.tfh-wildau.de/~pboettch/