PCI bus drivers, i2c-dev

[Mark D. Studebaker]

Well I'm a little slow :) but I fixed i2c-dev.

And checked in i2c-viapro backported from 2.5.69.
I'll try and backport and/or fix more of the PCI bus drivers that are obviously broken:
	i2c-ali1535
	i2c-alil15x3
	i2c-i801
	i2c-i810
	i2c-piix4
	i2c-sis5595

Mark D. Studebaker wrote:
> Most of the PCI bus drivers don't work. Sigh.
> Kyosti got them mostly converted to the PCI module style
> but then didn't fill in the pci_device_id table.
> 
> I think the best thing to get them working is to backport them from 2.5
> rather than finish what Kyosti started, let's use what Greg did.
> The only real problem is the dev_dbg, dev_info, etc.
> I'm making a kernel/include/i2c_compat.h to define these to make it easier.
> Is that OK?
> 
> I was testing things with i2cdetect and i2cdump but they just oops.
> Thought it was the PCI driver problem but now I think it's in
> i2c-dev.c in i2cdev_open(). I think it's
> __MOD_INC_USE_COUNT(client.>adapter->owner) that is oopsing? (null 
> pointer).
> 
> Every time it oopses the refcount on i2c-dev module increments.
> Is that what it's supposed to do? Or is it the adapter module? Both?
> Of course when it increments and then oopses I have to reboot to
> try and fix it.
> 
> So it's slow going here. Any ideas? Are i2cdump/i2cdetect working for 2.5?
> 
>