Date: Sun, 21 Oct 2007 19:53:06 -0700 From: Sam Leffler <sam@errno.com> To: "Alexandre \"Sunny\" Kovalenko" <alex.kovalenko@verizon.net> Cc: freebsd-mobile@freebsd.org Subject: Re: Problems running ath with bgscan enabled Message-ID: <471C1092.1030209@errno.com> In-Reply-To: <1193017503.844.14.camel@RabbitsDen> References: <20071021003032.D2DA04500E@ptavv.es.net> <471AC14A.9060806@errno.com> <1193017503.844.14.camel@RabbitsDen>
next in thread | previous in thread | raw e-mail | index | archive | help
Alexandre "Sunny" Kovalenko wrote: > On Sat, 2007-10-20 at 20:02 -0700, Sam Leffler wrote: >> Kevin Oberman wrote: >>> Ever since the update of the HAL and the new 802.11 code a few months >>> ago (in current), I have been having problems, especially when I was in >>> locations with multiple APs available. >>> >>> The problem was periodic loss of association. If I was at home where I >>> can usually only see a couple of APs, the problem would occur an >>> average of about 10 times a day with re-association in about 10 seconds, >>> sometimes less. This is annoying, but not very disruptive. >>> >>> When away from home and in an urban location where there are several >>> APs, both the frequency of the loss of connectivity and the time to >>> re-associate increase. Last week at the NANOG meeting there were 4 >>> conference b/g APs and 4 11a APS seen at all times along with the >>> hotel's tmobile and a couple of random ones. Only the two conference >>> SSIDs (nanog-arin and nanog-arin-a) were in my wpa_supplicant.conf >>> file. >>> >>> My bgscan interval was the default 300 and almost every 5 minutes I >>> would lose connectivity, often for over a minute. When I made the >>> association between the bgscan interval and the drop-outs, I did an >>> "ifconfig ath0 -bgscan" and things stabilized immediately. No further >>> problems. >>> >>> ath_hal: 0.9.20.3 (AR5210, AR5211, AR5212, RF5111, RF5112, RF2413, RF5413) >>> ath0: <Atheros 5212> mem 0xb4000000-0xb400ffff irq 11 at device 2.0 on pci11 >>> ath0: mac 5.9 phy 4.3 radio 3.6 >>> >>> This is really not acceptable for default behavior as it won't be clear >>> to many how to fix it. Is it expected or is something wrong? >>> >>> Thanks! >> There hasn't been a hal change in a long time. bgscan works by telling >> the ap your station is in power save mode, leaving the channel briefly, >> and then returning. I have not seen any issues with it w/ or w/o >> wpa_supplicant. >> >> Maybe you can provide some information to help diagnose this? I see >> nothing helpful like a log from wpa_supplicant or statistics. >> >> Sam >> _______________________________________________ >> freebsd-mobile@freebsd.org mailing list >> http://lists.freebsd.org/mailman/listinfo/freebsd-mobile >> To unsubscribe, send any mail to "freebsd-mobile-unsubscribe@freebsd.org" > This is "me too" and the question. The question is: where did sysctl > dev.ath.0.debug go? I have defined ATH_DEBUG in if_ath.c and rebuilt the > module, but id did not seem to appear. I understand that log, produced > with that sysctl set to 0x00040000, would be the first thing you need to > look into the issue, so if you would nudge me in the proper direction, I > will be happy to produce it. I have two stations in my house and > experience the problem, Kevin describes, on the regular basis. Turning > off bgscan helped enormously (thank you, Kevin). > ath debug msgs are at the wrong level, you want things at the net80211 layer. I can't tell you why you're having problems turning on ath debug msgs, I do it all the time either by setting ATH_DEBUG in a kernel config file or in the module Makefile. Sam
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?471C1092.1030209>