Aircrack - Fast Wep Cracking Tool, got it built, need to start testing Options

[Foxdie]

*BUMP* Please can someone get this tool updated or give us a resolution on how to get it working?

Same old malloc(80) error for me as well after installing on C860 / pdaXrom 1.1.0 RC8

[silvio]

I have tested aircrack only under cacko 1.22 lite.
Maybe it is incompatible with pdaXrom.

I have unchecked pdaXrom in ELSI during upload.

For pdaXrom it should be better to recompile this package because it could be significant faster (gcc 3).

[jfv]

I am using the Sharp ROM and I get this error too, as reported earlier. Silvio, could you download the file from ELSI and install in your Zaurus and see if it works? Maybe the uploaded file is different from what you have on your machine.

Thanks,

Felipe

[stupkid]

I am also getting the malloc error. Looking at the aircrack binary Silvio and I should have the exact same libraries/hardware/OS. Silvio, can you crack wellenrieter capture files? If so, what are the exact commandline arguments that you use?

I wonder if the issue is that I don't have enough packets to get aircrack to work properly. You need like 500,000 packets to crack 128-bit keys. I certainly have not captured that many packets yet. Hmm, some more experimentation is in order.

[jfv]

I downloaded the source from the original site and compiled it on the cluster at handhelds.org. The binary of different size as the binary from ELSI. I haven't got my Zaurus with me (shame on me) so I can't test it. I'll test it tonight, but if anyone wants to try it, I'll attach it here.

Felipe

p.s. I can't seem to attach it so here it is.

[berkenb]

I looked at the aircrack source a while back. I believe the way it works (and I don't really know anything about it, so I might just be talking nonsense here) is by gathering statistical information about 24bit IVs that are part of each wireless packet.
In order to do this, aircrack allocates 5 bytes of memory for each possible IV - i.e. 5*2^24 = 80MB. Hence the program tries to allocate one big 80MB chunk of memory (cf. row 1012 in aircrack.c, version 2.1).
Since the Z has (at most) 64MB of ram (so your free memory is considerably less than that), you will need a big swap file in order to get this to work, otherwise it will always fail. So that's the error message you guys see.
Besides, I think that a capture file with enough packets in it will easily run in the hundreds of MB...
It seems like this whole endeavor is a little impractical on the Z.

[stupkid]

I downloaded the source from the original site and compiled it on the cluster at handhelds.org. The binary of different size as the binary from ELSI. I haven't got my Zaurus with me (shame on me) so I can't test it. I'll test it tonight, but if anyone wants to try it, I'll attach it here.

Felipe

p.s. I can't seem to attach it so here it is.

FYI I have the exact same malloc issue as with Silvio's aircrack binary.

[berkenb]

Maybe my last answer was a little too long winded....
I think there is absolutely nothing wrong with the binaries you are trying - it is just that aircrack needs to be able to allocate 80MB of memory, and that is impossible on any Z to date (even the newer ones only have 64megs of memory), unless you have a large swap file somewhere...

[charlesa]

Maybe my last answer was a little too long winded....
I think there is absolutely nothing wrong with the binaries you are trying - it is just that aircrack needs to be able to allocate 80MB of memory, and that is impossible on any Z to date (even the newer ones only have 64megs of memory), unless you have a large swap file somewhere...

OK, to test this what is the best way to set a swap file size on a SD card?

[berkenb]

Maybe my last answer was a little too long winded....
I think there is absolutely nothing wrong with the binaries you are trying - it is just that aircrack needs to be able to allocate 80MB of memory, and that is impossible on any Z to date (even the newer ones only have 64megs of memory), unless you have a large swap file somewhere...

OK, to test this what is the best way to set a swap file size on a SD card?

I am not the world's leading expert in doing this, but in order to create a swap file on your SD card, you could follow something like the following steps:

CODE

dd if=/dev/zero of=/mnt/card/swapfile bs=1M count=64
mkswap /mnt/card/swapfile
swapon /mnt/card/swapfile

This creates a 64mb file called "swapfile" on /mnt/card containing just 0s, initializes it as a swapfile, and then turns the swapfile on. As a more permanent solution, you would add an entry for this swapfile in your /etc/fstab, but the above steps will do as a quick and dirty method. You can check the status with

CODE

cat /proc/swaps

and turn it off with

CODE

swapoff /mnt/card/swapfile

Mind you though that swapping on SD is probably agonizingly slow and puts some wear on your card (flash memory doesn't have the same amount of write cycles a harddrive has). I think I have used a swapfile on SD before, but only as a test, and certainly never for extended periods of time.
Hope this helps...

[jfv]

I created the swapfile (64MB) on my SD card and aircrack did run, although it said my dump file did not contain enough data to recover the key
I won't leave the swapfile there, takes too much room and, as mentioned above, there are some drawbacks. But it's good to know that, in a pinch, I can make it work.

Felipe

[undrwater]

Confirmed.

I set up a swap partition on my CF HD, and it now works. Not sure why swapd didn,t work on my SD (actually I think it did something bad to it).

Cool

[eji]

When I've deleted some MP3s and have a bit more space, I'll try creating a swapfile to see if it works.

FYI, maslovsky's memory applet has a handy GUI for creating swapfiles of any size to any media.

[speculatrix]

here's my theory:

when creating swap files on memory cards, I would recommend the larger the swap file the better!

you want to spread the wear on the memory card as much as possible, if you force the kernel to use the smallest swap file possible, it will be writing the same set of memory cells intensely (assuming that the card can't somehow write new data to a completely different region of flash), but if you make the swap file really huge, it will not need to use the same area twice?

how about swapping over the network? on a previous project, LinuxAP, using a eumitcom (x86-compatible) system, it was possible to swap over the network block device, which could make a big difference to performance (it only had 4MB of ram).

Paul

[stupkid]

Hmm, a low memory version aircrack would be nice.