b43 wifi taking forever to find access points 3.8.13-gentoo

eccerr0r

Odd...

I was running 3.7.10-gentoo-r1 x86 and upgraded to 3.8.13-gentoo.

The new kernel through Networkmanager takes 20 minutes before it detects an access point (Hardware: HP Pavillion laptop with b43 driver for wifi)! Also, "iwlist scan" reports no access points for that duration. Note that this appears to be access point specific and seemingly random... the behavior is unpredictable.

Booting back to 3.7.10-gentoo-r1 the problem disappears.

The main change I did was disable SMP support in the 3.8.13-gentoo kernel, as the 3.7.10-gentoo-r1 kernel I had was supposed to be a general purpose kernel - the laptop will never get a SMP CPU so I figure I should disable it.

Anyone seeing this kind of behavior?

06:05.0 Network controller: Broadcom Corporation BCM4318 [AirForce One 54g] 802.11g Wireless LAN Controller (rev 02)

666threesixes666 · Posted: Fri Jun 07, 2013 8:15 pm Post subject:

yes i see this behavior on my laptop, i have to iwlist scan like 8-15 times to get an AP to load. disabling SMP is an unacceptable work around. 7 dollar ath9k usb dongles on ebay is my solution, that's in the mail yet... i need in kernel support so i don't have to tinker with downloading firmware. add usb extenders, and pringles can... and the 13 dollar cantenna is a completely reasonable work around for me. i'm blacklisting and boycotting anything other than ath devices, same for anyone else consulting me on hardware that's worth supporting.

eccerr0r · Posted: Fri Jun 07, 2013 9:36 pm Post subject:

Ah interesting. Also have b43? I don't know if SMP was the ultimate reason but that was the only thing I knew I changed when going from 3.7.10 to 3.8.13. Unfortunately this machine is now 1500 miles away and I need to do remote debug...

I haven't yet seen the problem on my iwlagn. I have not updated my ath5k to 3.8.13, hoping it won't see the problem either.
_________________
Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching?

666threesixes666 · Posted: Fri Jun 07, 2013 9:58 pm Post subject:

my laptop rejected 3.7 and 3.8 series of kernels, i only noticed it when 3.9.4 compiled. didn't notice in 3.6.7. kernel regression. ill have to drop the old 3.4 stable in on that machine and test that.

TomWij · Retired Dev Joined: 04 Jul 2012 Posts: 1553

Please file bugs at https://bugs.gentoo.org and check if https://bugs.gentoo.org/show_bug.cgi?id=466920 may be relevant to what you experience.

_______0 · Guru Joined: 15 Oct 2012 Posts: 521

more modern adapters are being sold for a pittance.

eccerr0r · Posted: Sat Jun 08, 2013 4:48 pm Post subject:

_______0 · Guru Joined: 15 Oct 2012 Posts: 521

eccerr0r · Posted: Sat Jun 08, 2013 6:09 pm Post subject:

The cost is higher if having it completely hidden inside a machine has higher value.
Plus the fact those silly mini-PCI/firmware have locking for FCC compliance.
_________________
Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching?

666threesixes666 · Posted: Sat Jun 08, 2013 9:16 pm Post subject:

i had an ath9k minipci chip in my hand, my laptop rejected it and would not boot with a strange chip in it. cheap chips sink ships. usb dongles and usb extenders and pringles cans are a less than 10 dollar solution that will get you in kernel support and support for long range wifi. getting the wallet out is not a helpful solution, but downloading the firmware, and banging your head against the wall is not productive either.

eccerr0r · Posted: Sun Jun 09, 2013 12:31 am Post subject:

Well, this is not my machine, this is my parents' machine and well, it's also 1500 miles away. I can slowly experiment with kernel upgrades but getting a USB adapter is hard enough, getting them to plug it in and try to get it to work is an even tougher problem (plus the fact not having the wireless up means no remote debug...)

But yeah the minipci problem sucks. They (the HW manufacturer and minipci mfrs) tend to lock them to their respective machines to prevent people from breaking FCC rules in laptops. It's unfortunate too that this is actually a kernel regression or at least unexpected behavior from a patch...

I just hope I can get this "fixed" without completely breaking it. Remote debug is very dicey on network links...
_________________
Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching?

666threesixes666 · Posted: Sun Jun 09, 2013 1:46 am Post subject:

id just roll the kernel back and wait for it to be fixed in that case. remove the 3.8.13 from the boot selections.

eccerr0r · Posted: Tue Jun 11, 2013 11:14 pm Post subject:

This thread may be the same issue: https://forums.gentoo.org/viewtopic-t-960664-highlight-.html

I also heard that warm rebooting "fixes" the problem, but still seems like a regression... Need to test if this works or not.

blast this broadcom.
_________________
Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching?