On 06/01/15 10:13, intrigeri wrote:
> Hi,
> 
> David Arnstein wrote (03 Jan 2015 19:28:01 GMT) :
>> Today I downloaded a new ISO named
>> tails-i386-feature_jessie-1.3-20150103T1234Z-67d9850.iso . I burned it to a 
>> DVD and
>> I booted my Macbook Air from this DVD.
> 
>> The WiFi function of the machine seems to be working, up to a point. The 
>> Tails GUI
>> presented all of the WiFi SSIDs in my vicinity. I tried my own SSID but it 
>> failed
>> to connect.
> 
>> /var/log/syslog contains many interesting messages so I attach it to this 
>> email.
> 
>> I don't know how to debug this issue further, but if you would send detailed
>> instructions I should be able to follow.
> 
> Thanks! I see MAC spoofing failures in the logs:
> 
>   Jan 3 19:14:12 localhost kernel: [   62.201301] wl0: online cpus 1
>   Jan 3 19:14:12 localhost kernel: [   62.201709] wlan0: Broadcom BCM43a0 
> 802.11 Hybrid Wireless Controller 6.30.223.248 (r487574)
>   Jan 3 19:14:12 localhost kernel: [   62.441758] wl0: wl_set_mac_address: 
> error setting MAC addr override
>   Jan 3 19:15:36 localhost spoof-mac: Trying to spoof MAC address of NIC 
> wlan1...
>   Jan 3 19:15:36 localhost spoof-mac: macchanger failed for NIC wlan1, 
> returned 0 and
>   said: [ERROR] Could not change MAC: interface up or insufficient 
> permissions: Too
>   many open files in system#012Current MAC:  48:d7:05:a1:d8:73 
> (unknown)#012Permanent
>   MAC: 48:d7:05:c3:d3:81 (unknown)
>   Jan 3 19:15:36 localhost kernel: [  151.086135] wl0: wl_set_mac_address: 
> error setting MAC addr override

It's interesting that the "Current MAC" reported actually is different
than the "Permanent MAC", which otherwise would indicate that MAC
spoofing indeed succeeded. Could it be that the driver's/linux' state of
the MAC address *is* changed (and that's what's reported), but that it
remains unchanged in the hardware, so that their states are different?
If so I think the least we should do to prevent the real MAC address
from leaking is to treat errors from macchanger more seriously than we
currently do (see #8571).

Other than that, the (p)error "Too many open files in system" doesn't
make much sense to me. Probably it's just that the (notoriously crappy)
broadcom driver returns an incorrect error code. In fact, I had a quick
look and the error code returned from wl_set_mac_address() is in turn
returned by some function defined in the broadcom driver's "enormous
binary blob [wlc_hybrid.o] which contains _all_ of the 802.11 protocol
stack presented as a monolithic ethernet driver using WEXT" [1]. From a
situation like that, I don't expect much consistency with the running
kernel, but I confess I'm no expert in these matters.

[1]
http://blog.gmane.org/gmane.linux.ubuntu.devel.kernel.general/month=20090501/page=7

> So,
> 
>   * David, may you please opt-out from MAC spoofing in Tails Greeter,
>     and retry?

This is crucial advice since a MAC spoofing failure should result in the
driver being unloaded (to prevent the device from leaking the real MAC
address) and a syslog entry about whether that succeeded or not.
However, there's no such log entry in your attached syslog, which is
very concerning, and that you actually got a list of SSIDs is even more
so. After some digging the explanation is pretty simple: see #8571. Yikes!

Note that the above issue (#8571) isn't really relevant for you, David.
Like intrigeri suggested, please try again with MAC spoofing disabled in
Tails Greeter. It could be that MAC address spoofing is terribly broken
in the broadcom driver and sets the device in a bad state, and this is
what you are experiencing.

Cheers!

_______________________________________________
tails-support mailing list
[email protected]
https://mailman.boum.org/listinfo/tails-support
To unsubscribe from this list, send an empty email to 
[email protected].

Reply via email to