.. | ||
_ipk-building | ||
impersonate-lge.com.rb | ||
json-list.rb | ||
README.md | ||
webos.paw |
LG webOS
TV
name | value |
---|---|
model | 43UH6100 |
product | 3.0 |
firmware | 4.30.40 |
features | app marketplace, live TV listings |
vulnerabilities | all phone-home calls are done over HTTP |
the 43UH6100
is a 'smart' TV, running LG's webOS
since it is a fair assumption it is running OpenWrt underneath, the original goal
was rooting the device, but initial investigations showed some other interesting vectors.
digging
nmap
from nmap -PN -sV <device
, we get:
PORT STATE SERVICE VERSION
1175/tcp open upnp
3000/tcp open http LG smart TV http service
3001/tcp open ssl/http LG smart TV http service
9998/tcp open http Google Chromecast httpd
aside from the obvious flag running of both HTTP and HTTPS versions of (likely) the same service, interested to see that the Chromecast plugged in to the TV is also being exposed on the same IP as the TV.
since there is an LG smart TV app available for Android/iOS, assuming that there is an API of some sort running on 3000
or 3001
, so:
$ curl http://<device>:3000
Hello world
we see the same response on 3001
, but have to use -k
as the device uses a self-signed certificate.
so, something is there, we just don't know how to talk to it yet.
sniffing
switching tactics and connected the TV to a wireless network that has a tap, and we start to see some interesting things:
on boot
every time the TV starts up, within 30 seconds, it calls home:
POST /CheckSWAutoUpdate.laf HTTP/1.1
Accept: */*
User-Agent: User-Agent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1)
Host: snu.lge.com:80
Connection: Keep-Alive
Content-type: application/x-www-form-urlencoded
Content-Length: 572
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
HTTP/1.1 200 OK
Date: Wed, 16 Nov 2016 08:23:56 GMT
Content-length: 508
Content-type: application/octet-stream;charset=UTF-8
Pragma: no-cache;
Expires: -1;
Content-Transfer-Encoding: binary;
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
that looks a lot like base64 encoded data, and when decoded, yields
request:
<REQUEST>
<PRODUCT_NM>webOSTV 3.0</PRODUCT_NM>
<MODEL_NM>HE_DTV_W16P_AFADATAA</MODEL_NM>
<SW_TYPE>FIRMWARE</SW_TYPE>
<MAJOR_VER>04</MAJOR_VER>
<MINOR_VER>30.40</MINOR_VER>
<COUNTRY>US2</COUNTRY>
<COUNTRY_GROUP>US</COUNTRY_GROUP>
<DEVICE_ID>de:ad:be:ef:ca:fe</DEVICE_ID>
<AUTH_FLAG>N</AUTH_FLAG>
<IGNORE_DISABLE>N</IGNORE_DISABLE>
<ECO_INFO>01</ECO_INFO>
<CONFIG_KEY>00</CONFIG_KEY>
<LANGUAGE_CODE>en-US</LANGUAGE_CODE>
</REQUEST>
pretty standard, but the auth_flag
, ignore_disable
and config_key
values are potentially interesting
response:
<RESPONSE>
<RESULT_CD>900</RESULT_CD>
<MSG>Success</MSG>
<REQ_ID>00000000000000000001</REQ_ID>
<IMAGE_URL></IMAGE_URL>
<IMAGE_SIZE></IMAGE_SIZE>
<IMAGE_NAME></IMAGE_NAME>
<UPDATE_MAJOR_VER></UPDATE_MAJOR_VER>
<UPDATE_MINOR_VER></UPDATE_MINOR_VER>
<FORCE_FLAG></FORCE_FLAG>
<KE></KE>
<GMT>16 Nov 2016 08:23:56 GMT</GMT>
<ECO_INFO>01</ECO_INFO>
<CDN_URL></CDN_URL>
<CONTENTS></CONTENTS>
</RESPONSE>
much more interesting than the request:
key | assumption |
---|---|
IMAGE_URL |
the URL of a firmware update |
IMAGE_SIZE |
the size of the firmware update - are they doing this instead of checksum? |
IMAGE_NAME |
the name of the firmware update - not sure why this is necessary |
UPDATE_MAJOR_VER |
the major version of the firmware update |
UPDATE_MINOR_VER |
the minor version of the firmware update |
FORCE_FLAG |
whether or not to force the update - unclear if true |
CDN_URL |
URL that the firmware update is available at |
CONTENTS |
none |
half an hour of playing around with both the input and output here didn't yield any immediate results, but there is definite potential.
to speed this along, observe a session where the TV updated its firmware from the manufacturer
channel search
foo
application marketplace
bar
impersonating
baz
channel guide
barney
application update
fizzbang