IoT Inspector Research Lab Security Advisory IOT-20210414-0  
title: Cisco RV series Authentication Bypass and Remote Command   
vendor/product: Cisco (  
vulnerable version: RV16X/RV26X: & below.  
RV34X: & below.  
fixed version: RV16X/RV26X:   
CVE number: CVE-2021-1472, CVE-2021-1473  
impact: 5.3 (medium) CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:L/A:N  
8.8 (high) CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:L  
reported: 2021-01-02  
publication: 2021-04-14  
by: T Shiomitsu, IoT Inspector Research Lab  
Vendor description:  
The RV series devices are Cisco's line of small business routers with extra  
functionality, including VPN and other security measures.   
Vulnerability overview/description:  
All Cisco RV-series routers suffer from an authentication bypass vulnerability.  
The RV34X series are also affected by a command injection vulnerability in the  
sessionid cookie, when requesting the /upload endpoint. A combination of these  
issues would allow any person who is able to communicate with the web   
interface to run arbitrary system commands on the router as the www-data user.  
Root Cause Analysis:  
CVE-2021-1472: /upload Authentication Bypass Vulnerability  
While Cisco has noted that this issue also affects the RV160, I will provide  
a RCA for only the RV34X series here.  
The RV340 web interface is served by nginx on port 443. The nginx configuration  
(found in files in /etc/nginx) is such that requests made to the web interface  
URIs /upload, /form-file-upload and   
/api/operations/ciscosb-file:form-file-upload are all proxied to a CGI binary   
called upload.cgi. Depending on which URI is requested, the behaviour of the   
binary will be slightly different.  
While some attempt was introduced in recent firmware revisions to prevent   
unauthenticated access to the functionality available at the /upload endpoint,  
the authentication check is incomplete. An attacker simply has to pass any   
generic Authorization header as part of the request to bypass the authorization  
check. This can be seen in web.upload.conf:  
location /upload {  
set $deny 1;  
if ($http_authorization != "") {  
set $deny "0";  
if (-f /tmp/websession/token/$cookie_sessionid) {  
set $deny "0";  
if ($deny = "1") {  
return 403;  
As can be seen, the $deny is set to 0 if the $cookie_sessionid is valid (i.e.   
that the authorization file exists on the system). But it also set to 0 if the  
$http_authorization value (i.e. the Authorization header) is not blank.   
Therefore, passing any value to an Authorization header can allow an attacker  
access to the /upload endpoint.  
CVE-2021-1473: /upload sessionid Command Injection Remote Code Execution  
Within the main() function in upload.cgi, the HTTP_COOKIE environmental   
variable is read, and the value from the sessionid cookie is extracted using   
a simple series of strtok_r and strstr. This specific sessionid-reading logic   
is notable because, due to the strtok_r call, it's not possible to use ";"   
characters in any injection, as it will prematurely terminate the injection   
string. In pseudocode, it looks like this:  
if (HTTP_COOKIE != (char *)0x0) {   
cookie = StrBufToStr(cookie);   
cookie = strtok_r(cookie, ";", &saveptr);   
while (cookie != 0x0) {   
cookie = strstr(cookie, "sessionid=");   
if (cookie != 0x0) {   
sessionid_cookie_value = pathparam_ + 10;   
Because our HTTP request is made to the /upload URI, the main() function in   
upload.cgi calls a function at 000124a4, which I've named handle_upload().   
This function takes a pointer to the sessionid cookie value as its first   
void handle_upload(char *sessionId, char *destination, char *option,   
char *pathparam, char *fileparam, char *cert_name, char *cert_type,   
char *password)   
It also takes several other arguments, each of which are populated by the   
multipart request parsing that takes place in the main() function. The names   
I've given these arguments roughly align with the names of the parameters   
that this multipart ingesting logic looks for.  
(Depending on what string is passed as the pathparam parameter, slightly   
different code paths will be taken, which means that slightly different checks  
must be bypassed to be able to reach the vulnerable code. In this example, I   
am using a request with the pathparam set to "Configuration", so the pseudocode  
I'm showing reflects this.)  
Within handle_upload(), a curl command is constructed with a call to sprintf,   
the resulting buffer of which is then passed directly to popen:  
ret = strcmp(pathparam, "Configuration");   
if (ret == 0) {   
config_json = upload_Configuration_json(destination,fileparam);   
if (config_json != 0) {   
post_data = json_object_to_json_string(config_json);   
sprintf(command_buf, "curl %s --cookie \'sessionid=%s\' -X POST -H \'Content-Type: application/json\' -d\'%s\' ", jsonrpc_cgi, sessionId , post_data);   
__stream = popen(command_buf, "r");   
if (__stream != (FILE *)0x0) {   
The sessionid cookie value that we have passed in our request is passed   
directly into this sprintf() call. With a crafted sessionid value, we would   
therefore be able to inject arbitrary commands into this command buffer. This   
will run the command with the privileges of the upload.cgi process which, in   
this case, is www-data.  
Vulnerable / tested versions:  
Cisco RV16X, RV26X and RV34X series devices.  
Apply Cisco-supplied patch. For RV16X/26X, For RV34X,  
Advisory URL:  
Vendor contact timeline:  
2021-01-02: Initial disclosure made to Cisco PSIRT.  
2021-01-07: Confirmation of receipt of disclosure from Cisco PSIRT.  
2021-01-27: Confirmation that issue is valid from Cisco PSIRT.  
2021-02-12: Update from Cisco PSIRT.  
2021-03-23: We contact Cisco PSIRT for timeline update and CVE IDs.  
2021-03-23: Cisco PSIRT respond giving us timeline and CVE IDs.  
2021-04-07: Cisco release advisory.  
The IoT Inspector Research Lab is an integrated part of IoT Inspector.  
IoT Inspector is a platform for automated security analysis and compliance   
checks of IoT firmware. Our mission is to secure the Internet of Things. In   
order to discover vulnerabilities and vulnerability patterns within IoT devices  
and to further enhance automated identification that allows for scalable   
detection within IoT Inspector, we conduct excessive security research in the  
area of IoT.   
Whenever the IoT Inspector Research Lab discovers vulnerabilities in IoT   
firmware, we aim to responsibly disclose relevant information to the vendor  
of the affected IoT device as well as the general public in a way that   
minimizes potential harm and encourages further security analyses of IoT   
You can find our responsible disclosure policy here:  
Interested in using IoT Inspector for your research or product?  
Mail: research at iot-inspector dot com  
EOF T Shiomitsu / @2021