Quoting Jeff Garzik <jeff@xxxxxxxxxx>: > It's a highly standardized interface that provides information that's > either impossible or highly difficult to obtain elsewhere. For what it's worth, ifrename in the forthcoming Wireless Tools 29 will support reading symlinks on sysfs, which would allow matching the device and the driver name. Other userspace tools can do the same. $ readlink /sys/class/net/eth0/device ../../../devices/pci0000:00/0000:00:0d.0/0000:01:02.0 $ readlink /sys/class/net/eth0/device/driver ../../../../bus/pci/drivers/orinoco_plx That's 2.6.20, things may be slightly different on 2.6.21+ > If you are a userland process querying a network interface, that's the > only way to know which driver is attached, or the only way to build an > association between a PCI device and a network interface. > > So NAK this change. All network drivers should implement GDRVINFO, even > if they are not strictly ethernet drivers. I would say it's not yet time to remove it. But if a driver doesn't implement it (that's the case for bcm43xx_d80211), I don't think we should bother implementing it. -- Regards, Pavel Roskin - To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
