wpa_supplicant segmentation fault
Tue Feb 24 10:56:13 PST 2009
On Tue, Feb 24, 2009 at 04:42:52PM +0200, Cristian NAVALICI wrote:
> I'm struggling to get up an wireless card (EDIMAX PCI adapter based on
> Ralink RT2561/RT61 802.11g chipset). I managed to install the card
> drivers, everything seems to be ok.
Which drivers are you using?
> wpa_supplicant v0.6.8
> wpa_supplicant -Dralink -iwlan0 -c/etc/wpa_supplicant/wpa_supplicant.conf -dd
Are you sure you need to use the driver specific wrapper (-Dralink)?
Have you tried with the more generic WEXT one (-Dwext)?
> Segmentation fault
> More debugging info:
> (gdb) run /usr/local/sbin/wpa_supplicant -Dralink -iwlan0 -c/etc/wpa_supplicant/wpa_supplicant.conf
> Starting program: /usr/local/sbin/wpa_supplicant /usr/local/sbin/wpa_supplicant -Dralink -iwlan0 -c/etc/wpa_supplicant/wpa_supplicant.conf
> Program received signal SIGSEGV, Segmentation fault.
> 0x00002b24e2be647b in memcpy () from /lib64/libc.so.6
Could you please do the same but with backtrace showing the called
functions ('bt' in gdb)?
> This wpa_supplicant is built from the sources, but I tried either with
> I run Centos 5.2 64bit version.
> As a curiosity, on the same computer, but with Centos 5 32bit version,
> this seems to work without any problem. So I guess it's related to 64bit
It is possible that the data structure used for the scan results would
not have matching size in the kernel code and wpa_supplicant in some of
the configurations.. It looks like wpa_driver_ralink_get_scan_results()
does not check the values received from the driver, so it might easily
end up trying to copy huge memory areas if there is a mismatch in size
of alignment of the fields.
Jouni Malinen PGP id EFC895FA
More information about the Hostap