Hi all and let me repost my previous question again (with a few more updates, also directly to Linville, as well as to the list) as I STILL got no answer for it (Don't ignore it this time!):
I have a strange problem with my Asus WL-138G V2 (which is AFAIK a BCM4318 non-E card). It works fine with bcm43xx (since 2.6.22, works with 2.6.24's bcm43xx too), but not with b43. I can scan without problems (iwlist scan works, knetworkmanager lists networks properly), monitor mode works with kismet (both on main interface and on a dedicated rfmon interface), but I can't associate (authenticate timeout). Radiotap injection also doesn't work, packetspammer doesn't complain about errors, but there is nothing actually sent out. Basically, anything that involves transmitting is broken, while receiving works fine. I have tried with the following kernel+firmware combos (all 2.6.24 kernels were opensuse-pae, while the 2.6.24.1 one was opensuse-default, tried 2.6.25-rc3 in both vanilla and opensuse-pae flavors since my previous post, neither worked): -kernel 2.6.24+firmware 4.80.53 "wl_apsta.o" -kernel 2.6.24+firmware 4.80.53 "wl_apsta_mimo.o" -kernel 2.6.24.1+firmware 4.80.53 "wl_apsta.o" -kernel 2.6.24.1+firmware 4.80.53 "wl_apsta_mimo.o" -kernel 2.6.24+compat-wireless-2.6+firmware 4.80.53 "wl_apsta.o" -kernel 2.6.24+compat-wireless-2.6+firmware 4.80.53 "wl_apsta_mimo.o" -kernel 2.6.24+compat-wireless-2.6+firmware 4.150.10.5 "wl_apsta_mimo.o" -kernel 2.6.25-rc3 (vanilla and opensuse-pae)+firmware 4.80.53 "wl_apsta.o" -a recent wireless-testing kernel with opensuse patches+firmware 4.80.53"wl_apsta.o" (I only used this kernel for one day, as it messed up my ReiserFS root partition, had to reiserfsck --fix-fixable to regain 18 gigabytes of free space that was somehow marked as used - maybe I got my config wrong, or the 2.6.25-rc opensuse reiser patches don't like wireless-testing). All produced the same error. (Strangely, the wireless-testing git kernel showed this error with a ZD1211B and an RT73 USB stick, not just with my Broadcom PCI, so it might be a mac80211 issue.) I will post my dmesg output if needed, but I can see no error messages from the driver in it (even though I have enabled b43 and mac80211 debugging in the compat-wireless config.mk file), only a message about "authentication timed out". The message is the same for unencrypted, WEP104 open and WPA-PSK. I haven't tested WPA-EAP and WPA2-EAP as I don't have the needed hardware (no RADIUS), but I can test WEP40 open, WEP shared (both 40 and 104), 802.1X, WPA2-PSK and mixed-mode (v1+v2) WPA-PSK. (It might be interesting that while bcm43xx reported by card power levels as 15dBm, 2.6.24(.1)'s b43 (just like 2.6.25-rc3's one) reports it as 27dBm (which is IMHO a bit high, as that would be 400mW, while the card has an output of about 20mW - again, bcm43xx reports it as 15dBm!), and compat-wireless-2.6's b43 reports 0dBm, yes, 0dBm, just like the wireless-testing git kernel. That's quite odd. But it appears that compat-wireless-2.6 is right about the transmit power. :-) Is this a known bug? What can I do about it? Thanks, .NetRolller 3D P.S. Please answer, I don't want to repost again! This time really! P.P.S. Don't ignore this message! I can't stress it enough, I need an answer! -- Vista: [V]iruses, [I]ntruders, [S]pyware, [T]rojans and [A]dware. :-)
_______________________________________________ Bcm43xx-dev mailing list [email protected] https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
