Re: PCI driver

[Rolf Eike Beer <eike-kernel@sf-tec.de>]

Am Montag, 10. Oktober 2005 14:48 schrieben Sie:
>Rolf Eike Beer wrote:
>>IIRC the call to pci_enable_device() must be the first thing you do. This
>> will do the things like assigning memory regions to the device and so on.
>
>I fixed this one
>
>>Returning 0 in error cases is just wrong. And you free the assignments even
>> in case of success AFAICS. Try the return I introduced above and see what
>> happens.
>
>I fixed this one too ..
>
>
>I have fixed most of the stuff, it is partly working, not ready yet as
>there are some more things to be added to  ..
>I have attached what i was working on.

If the kmalloc() fails in mantis_pci_probe() you don't call 
pci_disable_device(). And you should kzalloc() instead of kmalloc() and 
memset().

It looks like you never use "__u16 vendor_id;" and "__u16 device_id;" in 
struct mantis_pci.

Eike

Attachment: pgppKcboS1Zyq.pgp
Description: PGP signature