From owner-freebsd-current@FreeBSD.ORG Sat Aug 13 17:02:42 2005 Return-Path: X-Original-To: freebsd-current@freebsd.org Delivered-To: freebsd-current@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 5685616A41F for ; Sat, 13 Aug 2005 17:02:42 +0000 (GMT) (envelope-from sam@errno.com) Received: from ebb.errno.com (ebb.errno.com [66.127.85.87]) by mx1.FreeBSD.org (Postfix) with ESMTP id CA73C43D45 for ; Sat, 13 Aug 2005 17:02:40 +0000 (GMT) (envelope-from sam@errno.com) Received: from [66.127.85.91] ([66.127.85.91]) (authenticated bits=0) by ebb.errno.com (8.12.9/8.12.6) with ESMTP id j7DH2cBd080293 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Sat, 13 Aug 2005 10:02:39 -0700 (PDT) (envelope-from sam@errno.com) Message-ID: <42FE28F8.2080605@errno.com> Date: Sat, 13 Aug 2005 10:08:08 -0700 From: Sam Leffler User-Agent: Mozilla Thunderbird 1.0.2 (X11/20050327) X-Accept-Language: en-us, en MIME-Version: 1.0 To: Michal Mertl References: <1123946218.61712.85.camel@genius1.i.cz> In-Reply-To: <1123946218.61712.85.camel@genius1.i.cz> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: Tai-hwa Liang , freebsd-current@freebsd.org Subject: Re: panic: bogus long slot station count 0 X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 13 Aug 2005 17:02:42 -0000 Michal Mertl wrote: > I'm getting easy to reproduce INVARIANTS panics on fresh CURRENT. > > The panicing machine has ath0 configured like this: "ifconfig ath0 > 192.168.0.1/24 media auto mode 11b mediaopt hostap ssid mig_ap_xx". > > When I associate to it with a notebook with an ipw card I get later > several panics with the same INVARIANTS cause - line 2073 in > src/sys/net80211/ieee80211_node.c. > > The "normal" panic is this scenario - a station connects and then leaves > (is turned off or something). > > I run '80211debug assoc' and this is what I get before the panic: > > ath0: [00:0c:f1:3e:8b:07] recv probe req > ath0: [00:0c:f1:3e:8b:07] station associated at aid 1: short preamble, > long slot time > ath0: [00:0c:f1:3e:8b:07] station with aid 1 leaves > > After the association was formed I put the ipw0 card 'down' and after > short while (a minute or so) I received the third message above and the > panic with the traceback below. I don't understand much what 'short > preamble' means but I'd expect to see 'S' character in 'ifconfig ath0 > list sta' command output on the AP if the station was really doing > shoart preamble. > > Output of 'ifconfig ath0 list sta' on the AP before the crash is this: > > ADDR AID CHAN RATE RSSI IDLE TXSEQ RXSEQ CAPS ERP > 00:0c:f1:3e:8b:07 1 1 1M 25 210 2 208 E 0 Sigh, guess we need to still reference ic_curmode instead of looking at the channel characteristics (this was a recent change to deal with your problem with switching mode from 11a->11g w/ a station associated). The problem is that ic_curmode needs to be completely removed. I've done this in p4 but the changes are too extensive to commit at this time. I'll backout the change for now. Separately, I recently learned that 11g ap mode w/ a mixed bss is broken. I forgot to update the ERP ie in the beacon frame so 11g stations are not informed when an 11b sta joins the bss and slot time and/or protection are altered. I won't get to fix this for a few weeks. Sam