Jim McCloskey wrote: > Hello. I am the owner of a card (a LinkSys Wireless-G WPC54GS (ver.2)) > which has a 4318 chip: > > kernel: bcm43xx: Chip ID 0x4318, rev 0x2kernel: bcm43xx: Number of cores: 4 > kernel: bcm43xx: Core 0: ID 0x800, rev 0xd, vendor 0x4243, enabled > kernel: bcm43xx: Core 1: ID 0x812, rev 0x9, vendor 0x4243, disabled > kernel: bcm43xx: Core 2: ID 0x804, rev 0xc, vendor 0x4243, enabled > kernel: bcm43xx: Core 3: ID 0x80d, rev 0x7, vendor 0x4243, enabled > > Last July I wrote to this list about the level of success I was having > with getting the card to work under Debian testing [1]. At that point, > the card was managing to associate with the access point and obtain an > IP address about one time in ten. That was with kernel version 2.6.17. > > I have barely used the card since then---because it wasn't really > usable, and because I scan or read the digest version of this list > every day and my impression from that was that the issues around the > 4318 chipsets hadn't really been addressed yet. > > However, three days ago, I thought I should try again. I've been using > the card with two slightly different kernel versions: 2.6.18.1 > hand-compiled from kernel.org and 2.6.18.3 packaged for Debian. > > And it's been working just fine. True, I can't use Debian's ifup > script, and automation utilities like network-manager and KDE's > kwifi-manager also can't create a working connection, but if I do: > > ifconfig wlan up > iwconfig essid ESSID-NAME > dhclient wlan > > all works consistently very well. That is, I have had no failures of > association at all over the past 3 days, despite more or less > continuous use and despite the fact that I have been deliberately > connecting and disconnecting, suspending and resuming, and so on to > try to test the card (I'm listening to an Irish radio stream with it > just now). Also: `iwlist wlan scan' used to regularly cause hard > lockups of the system, but I haven't been able to induce that > behaviour over the 3 day period either. > > The only change here is the kernel version---same laptop, same access > point, same number of access points in the vicinity. > > I haven't tried with 2.6.19, but obviously I should. > > So this is just a progress report and a note of thanks to you all for > your work. Obviously, I'm happy. But am I wrong to be also a little > bit puzzled?
Although we have not made any progress on the transmit strength problem with 4318, 4311 and 4312 chips, we have fixed a great number of bugs that led to lockups, and we continue to make advances. Kernel 2.6.19 is better than 2.6.18 and 2.6.20 will be even better. In addition, there is a lot of stuff waiting for 2.6.21, and so on. Larry _______________________________________________ Bcm43xx-dev mailing list [email protected] https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
