Celejar wrote: > > > # This file was automatically generated by the /lib/udev/write_net_rules > # program, probably run by the persistent-net-generator.rules rules file. > # > # You can modify it, as long as you keep each rule on a single line. > # MAC addresses must be written in lowercase. > > # PCI device 0x14e4:0x4318 (bcm43xx) > SUBSYSTEM=="net", DRIVERS=="?*", ATTRS{address}=="00:19:7d:06:a5:44", > NAME="eth0" > > # PCI device 0x14e4:0x170c (b44) > SUBSYSTEM=="net", DRIVERS=="?*", ATTRS{address}=="00:16:d4:5e:1e:9c", > NAME="eth1" > > # PCI device 0x168c:0x0013 (ath_pci) > SUBSYSTEM=="net", DRIVERS=="?*", ATTRS{address}=="00:40:f4:e5:07:7e", > ATTRS{type}=="1", NAME="ath0" > > The two Broadcom lines (wired and wireless cards builtin to my Acer > Aspire 3690-2672 laptop) don't contain the 'ATTRS{type}=="1"' clause, > while the Atheros line (for a PCMCIA card) does. Adding the clause to > the Broadcom wireless line fixed the problem, and it now seems to be > working perfectly (no screen flicker, which I had seen with bcm43xx), > so thanks. > > My questions: > > A) I have only a very basic understanding of udev; what does that > clause mean? I couldn't figure it out by googling or looking in the > basic udev docs. > > B) I'm running Debian Sid; I assume the installer wrote the builtin > cards' lines and the running system the subsequently inserted PCMCIA > line. Is this a bug? Do you know where I should report it? I have > never (until my addition of your clause) touched the ruls file. > > C) I wasn't seeing this device naming problem with my old kernels > (2.6.18 - 2.6.22) and bcm43xx / ndiswrapper. Is this a 2.6.24 thing, > or a b43 thing, or some combination thereof, or something else entirely?
It is a mac80211-introduced "problem". With bcm43xx-softmac, there is only a single device presented, and one does not care what type it is when it is renamed. With mac80211 (used by b43), at least two devices are available, wmaster0 with type==0x801, and wlan0 with type==1. Without the ATTRS{type==1} clause, wmaster=>eth1 and wlan0=>wlan0_rename. With the clause, only wlan0 is subject to renaming and all works OK. On some distros, merely deleting the rename rule gets the correct one written upon reboot. Sorry, I don't know which ones fit that category. I would expect that it depends on the version of udev that is included. If I were you, I would put it on Debian's bug-reporting system. Be sure that your message makes the problem associated with any wireless driver that uses mac80211. As I said before, the problem is more general than for b43. Larry _______________________________________________ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev