Arend,
I have just started testing brcmsmac with my BCM43224 card. I find it rather
unstable. A typical cycle between authentication and disassociating is as follows:
[ 1066.227729] wlan8: authenticate with 00:1a:70:46:ba:b1
[ 1066.238397] wlan8: send auth to 00:1a:70:46:ba:b1 (try 1/3)
[ 1066.440128] wlan8: send auth to 00:1a:70:46:ba:b1 (try 2/3)
[ 1066.443739] wlan8: authenticated
[ 1066.448139] wlan8: associate with 00:1a:70:46:ba:b1 (try 1/3)
[ 1066.456657] wlan8: RX AssocResp from 00:1a:70:46:ba:b1 (capab=0x411 status=0
aid=1)
[ 1066.456673] wlan8: associated
[ 1066.458051] ieee80211 phy0: brcms_ops_bss_info_changed: qos enabled: true
(implement)
[ 1066.458067] ieee80211 phy0: brcmsmac: brcms_ops_bss_info_changed: associated
[ 1066.458087] ieee80211 phy0: changing basic rates failed: -22
[ 1066.458101] ieee80211 phy0: brcms_ops_bss_info_changed: arp filtering:
enabled true, count 1 (implement)
[ 1076.464670] wlan8: disassociating from 00:1a:70:46:ba:b1 by local choice
(reason=3)
[ 1076.465981] ieee80211 phy0: brcms_ops_bss_info_changed: qos enabled: false
(implement)
[ 1076.466051] ieee80211 phy0: brcmsmac: brcms_ops_bss_info_changed: disassociated
[ 1076.466074] ieee80211 phy0: brcms_ops_bss_info_changed: arp filtering:
enabled false, count 1 (implement)
[ 1076.490224] cfg80211: Calling CRDA to update world regulatory domain
[ 1076.507338] wlan8: deauthenticating from 00:1a:70:46:ba:b1 by local choice
(reason=3)
The period of time that the interface is up ranges from 10 - 200 seconds. The
connection is WPA-TKIP and the kernel is master-2012-02-15-510-g20b79b4 from
wireless-testing with local patches applied to other drivers. Both brcmsmac and
bcma are untouched.
Thanks,
Larry
On 02/19/2012 09:53 PM, Larry Finger wrote:
> Arend,
>
> I have just started testing brcmsmac with my BCM43224 card. I find it rather
> unstable. A typical cycle between authentication and disassociating is as follows:
>
> [ 1066.227729] wlan8: authenticate with 00:1a:70:46:ba:b1
> [ 1066.238397] wlan8: send auth to 00:1a:70:46:ba:b1 (try 1/3)
> [ 1066.440128] wlan8: send auth to 00:1a:70:46:ba:b1 (try 2/3)
> [ 1066.443739] wlan8: authenticated
> [ 1066.448139] wlan8: associate with 00:1a:70:46:ba:b1 (try 1/3)
> [ 1066.456657] wlan8: RX AssocResp from 00:1a:70:46:ba:b1 (capab=0x411 status=0
> aid=1)
> [ 1066.456673] wlan8: associated
> [ 1066.458051] ieee80211 phy0: brcms_ops_bss_info_changed: qos enabled: true
> (implement)
> [ 1066.458067] ieee80211 phy0: brcmsmac: brcms_ops_bss_info_changed: associated
> [ 1066.458087] ieee80211 phy0: changing basic rates failed: -22
> [ 1066.458101] ieee80211 phy0: brcms_ops_bss_info_changed: arp filtering:
> enabled true, count 1 (implement)
> [ 1076.464670] wlan8: disassociating from 00:1a:70:46:ba:b1 by local choice
> (reason=3)
> [ 1076.465981] ieee80211 phy0: brcms_ops_bss_info_changed: qos enabled: false
> (implement)
> [ 1076.466051] ieee80211 phy0: brcmsmac: brcms_ops_bss_info_changed: disassociated
> [ 1076.466074] ieee80211 phy0: brcms_ops_bss_info_changed: arp filtering:
> enabled false, count 1 (implement)
> [ 1076.490224] cfg80211: Calling CRDA to update world regulatory domain
> [ 1076.507338] wlan8: deauthenticating from 00:1a:70:46:ba:b1 by local choice
> (reason=3)
>
> The period of time that the interface is up ranges from 10 - 200 seconds. The
> connection is WPA-TKIP and the kernel is master-2012-02-15-510-g20b79b4 from
> wireless-testing with local patches applied to other drivers. Both brcmsmac and
> bcma are untouched.
>
> Thanks,
>
> Larry
Hi Larry,
Thanks. There have been a couple of other people indicating a similar
issue. We run tests nightly on bcm43224 equipped systems and brcmsmac
mostly runs fine apart from occasional mac80211 interfacing issue, which
is one reason to run these tests. I will dig deeper into this.
Gr. AvS
On 02/19/2012 09:53 PM, Larry Finger wrote:
> Arend,
>
> I have just started testing brcmsmac with my BCM43224 card. I find it rather
> unstable. A typical cycle between authentication and disassociating is as follows:
>
> The period of time that the interface is up ranges from 10 - 200 seconds. The
> connection is WPA-TKIP and the kernel is master-2012-02-15-510-g20b79b4 from
> wireless-testing with local patches applied to other drivers. Both brcmsmac and
> bcma are untouched.
>
Hi Larry,
I tried reproducing this, but not luck so far. Could you provide me with
syslog including both kernel messages and wpa_supplicant. From the log
you provided it seems mac80211 triggers the disassoc. It could be a
beacon timeout, but I would like to be sure. More debug output from
mac80211 itself would also be helpful to understand the issue.
Gr. AvS
On 02/22/2012 10:14 AM, Arend van Spriel wrote:
>
> I tried reproducing this, but not luck so far. Could you provide me with syslog
> including both kernel messages and wpa_supplicant. From the log you provided it
> seems mac80211 triggers the disassoc. It could be a beacon timeout, but I would
> like to be sure. More debug output from mac80211 itself would also be helpful to
> understand the issue.
I may have misunderstood the symptoms with my system. The only AP that is
connected automatically is 802.11n; however, many of my cards now refuse to
connect to that AP. The BCM43224 is one of them. As a result, when the
connection to an 802.11g AP failed, the reconnection was to the N router, and it
failed.
I need to determine what is causing the failure to connect to the N router. It
might be a regression. After I find and fix that problem, then I can retest the
BCM43224.
Thanks, and sorry for the noise,
Larry