Re: [PATCH] deprecate pci_module_init

[Jeff Garzik]

Greg KH wrote:
> In short, pci_module_init() on 2.4 would return the number of pci
> devices bound to the device, on 2.6, it just always returns 0 if the
> driver was successfully registered, no knowledge of how many devices
> bound are ever returned.


Incorrect.  pci_register_driver() is the inconsistent one, as I've 
explained before (months ago).

pci_module_init() always returns 0 or an errno-based value, in 2.4 or 
2.6.  Thus is it the portable alternative.

Thus, changing drivers -away from- pci_module_init() makes them less 
portable, for zero apparent gain.  It's just a #define symbol at this 
point, leave it be.

The cost of "#define pci_module_init pci_register_driver" is zero, while 
the cost and impact of changing tons of drivers to the non-portable 
variant is non-zero.

	Jeff


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/