#390: Starting / shutting down hostapd causes kernel panic when non-AP VAPs
exist.
------------------------------------------------+---------------------------
Reporter: Stijn Tintel <[EMAIL PROTECTED]> | Owner:
Type: defect | Status: new
Priority: major | Milestone:
Component: madwifi: 802.11 stack | Version: trunk
Keywords: | Patch_attached: 0
------------------------------------------------+---------------------------
Starting or shutting down hostapd causes a kernel panic when there exist
VAPs that are running in a mode other then master/AP mode.
I captured the kernel output with netconsole:
{{{
Feb 10 01:47:19 192.168.50.6 unexpected operating mode 8
Feb 10 01:47:19 192.168.50.6 ------------[ cut here ]------------
Feb 10 01:47:19 192.168.50.6 kernel BUG at /root/drivers/madwifi-
ng/net80211/ieee80211_node.c:1860!
Feb 10 01:47:19 192.168.50.6 invalid operand: 0000 [#1]
Feb 10 01:47:19 192.168.50.6 Modules linked in:
Feb 10 01:47:19 192.168.50.6 aes
Feb 10 01:47:19 192.168.50.6 wlan_ccmp
Feb 10 01:47:19 192.168.50.6 wlan_tkip
Feb 10 01:47:19 192.168.50.6 wlan_xauth
Feb 10 01:47:19 192.168.50.6 sd_mod
Feb 10 01:47:19 192.168.50.6 sg
Feb 10 01:47:19 192.168.50.6 sr_mod
Feb 10 01:47:19 192.168.50.6 scsi_mod
Feb 10 01:47:19 192.168.50.6 ipv6
Feb 10 01:47:19 192.168.50.6 bridge
Feb 10 01:47:19 192.168.50.6 wlan_scan_ap
Feb 10 01:47:19 192.168.50.6 ide_cd
Feb 10 01:47:19 192.168.50.6 cdrom
Feb 10 01:47:19 192.168.50.6 parport_pc
Feb 10 01:47:19 192.168.50.6 parport
Feb 10 01:47:19 192.168.50.6 floppy
Feb 10 01:47:19 192.168.50.6 pcspkr
Feb 10 01:47:19 192.168.50.6 i2c_piix4
Feb 10 01:47:19 192.168.50.6 i2c_core
Feb 10 01:47:19 192.168.50.6 uhci_hcd
Feb 10 01:47:19 192.168.50.6 usbcore
Feb 10 01:47:19 192.168.50.6 shpchp
Feb 10 01:47:19 192.168.50.6 pci_hotplug
Feb 10 01:47:19 192.168.50.6 intel_agp
Feb 10 01:47:19 192.168.50.6 ath_pci
Feb 10 01:47:19 192.168.50.6 ath_rate_sample
Feb 10 01:47:19 192.168.50.6 wlan
Feb 10 01:47:19 192.168.50.6 ath_hal
Feb 10 01:47:19 192.168.50.6 agpgart
Feb 10 01:47:19 192.168.50.6 netconsole
Feb 10 01:47:19 192.168.50.6 e1000
Feb 10 01:47:19 192.168.50.6 rtc
Feb 10 01:47:19 192.168.50.6 ext3
Feb 10 01:47:19 192.168.50.6 jbd
Feb 10 01:47:19 192.168.50.6 mbcache
Feb 10 01:47:19 192.168.50.6 ide_disk
Feb 10 01:47:19 192.168.50.6 generic
Feb 10 01:47:19 192.168.50.6 ide_generic
Feb 10 01:47:19 192.168.50.6 piix
Feb 10 01:47:19 192.168.50.6 EIP is at ieee80211_node_leave+0x62/0x1a5
[wlan]
Feb 10 01:47:19 192.168.50.6 domlme+0x0/0x3f [wlan]
Feb 10 01:47:19 192.168.50.6 ieee80211_ioctl_setmlme+0x0/0x1c7 [wlan]
Feb 10 01:47:19 192.168.50.6 sys_ioctl+0x46/0x61
Feb 10 01:47:19 192.168.50.6
Feb 10 01:47:19 192.168.50.6 Kernel panic - not syncing: Fatal exception
in interrupt
Feb 10 01:47:19 192.168.50.6
Feb 10 01:47:19 192.168.50.6 Rebooting in 5 seconds..
}}}
In the first line of above output, the value for operating mode given is
always the number for the mode that the first non-AP VAP is running in
(monitor mode in this case).
I'm suspecting this to be a hostapd bug, but as Mike suggested on IRC, I'm
filing a ticket for it.
--
Ticket URL: <http://madwifi.org/ticket/390>
MadWifi <http://madwifi.org/>
Multiband Atheros Driver for Wireless Fidelity
