Bug#488267: Should add hostap modules
I should have expanded the argument in two ways. Building on:
I suppose this does make some kind of sense, as hostap_pci must in fact do (at least) two things in succession:
* create the master interface wifiN (where N=0 for us as it's the first such interface to be created)
* create an sta interface based on wifiN, and it chooses (hard-coded?) to call this wlanN
And if the installer used hostap_pci too, then presumably at install time, the device created would be wlan0.
we have:
(1) In fact, this must be what happens with madwifi too: it always
creates a sta interface athN, and I'm pretty sure there won't be a udev
rule in systems without madwifi that makes any reference to athN interfaces.
(2) And in fact you can see the process happening in the (installed
system) syslog extract I posted before:
Jun 28 08:08:46 obelix kernel: hostap_pci: Registered netdevice wifi0
Jun 28 08:08:46 obelix kernel: wifi0: Original COR value: 0x2
Jun 28 08:08:46 obelix kernel: prism2_hw_init: initialized in 204 ms
Jun 28 08:08:46 obelix kernel: wifi0: NIC: id=0x8013 v1.0.0
Jun 28 08:08:46 obelix kernel: wifi0: PRI: id=0x15 v1.1.0
Jun 28 08:08:46 obelix kernel: wifi0: STA: id=0x1f v1.4.3
Jun 28 08:08:46 obelix kernel: wifi0: defaulting to host-based encryption as a workaround for firmware bug in Host AP mode WEP
Jun 28 08:08:46 obelix kernel: wifi0: defaulting to bogus WDS frame as a workaround for firmware bug in Host AP mode WDS
Jun 28 08:08:46 obelix kernel: wifi0: Intersil Prism2.5 PCI: mem=0xe0500000, irq=9
Jun 28 08:08:46 obelix kernel: wifi0: registered netdevice wlan0
Jun 28 08:08:46 obelix kernel: orinoco 0.15 (David Gibson <hermes@gibson.dropbear.id.au>, Pavel Roskin <proski@gnu.org>, et al)
Jun 28 08:08:46 obelix kernel: orinoco_pci 0.15 (Pavel Roskin <proski@gnu.org>, David Gibson <hermes@gibson.dropbear.id.au> & Jean Tourrilhes <jt@hpl.hp.com>)
and a bit later, with nothing relevant in between
Jun 28 08:08:46 obelix kernel: udev: renamed network interface wifi0 to eth1
which implies that, in the presence of hostap_pci, (it or) orinoco_pci
has (silently??) created the interface wlan0.
Later in that syslog, udev acts:
Jun 28 08:08:46 obelix kernel: udev: renamed network interface wifi0 to eth1
and the next relevant thing we get is
Jun 28 08:08:46 obelix kernel: prism2: wlan0_rename: operating mode changed 3 -> 2
where something (hostap_pci?) has silently renamed eth1 to the ugly
wlan0_rename.
So arguably it was not a good idea (indeed - a bug?) for udev to be told
to rename wlan0 to eth1, and it would be better to stick with the
driver-mandated wlanN names.
Barry
Reply to: