MountLocker Ransomware


This is my report for a MountLocker Ransomware v5.0 sample, which is used by XingLocker ransomware group.

This ransomware uses a hybrid-cryptography scheme of RSA-2048 and ChaCha20 to encrypt files and protect its keys. Unlike other ransomware, MountLocker encrypts all of the ChaCha20 keys with a global ChaCha20 key before encrypting this global key with its RSA-2048 public key. The encrypted global key and the corresponding encrypted ChaCha20 key are appended at the end of each encrypted file.

This version includes a new worm feature that lets it self-propagate to other PCs on the network using IDirectorySearch and IWbemServices COM interfaces.

MountLocker has a sophisticated multithreading scheme, but its performance suffers from thread starvation due to recursive file traversal.

I won’t waste my time explaining why recursive file traversal is terrible anymore cause I have made my points through the last few reports. Please feel free to check out my Darkside analysis if you want to better understand the theory behind it!

alt text

Figure 1: XingLocker Ransomware leak site.


This v5.0 sample is a 64-bit .exe file.

MD5: 3808f21e56dede99bc914d90aeabe47a

SHA256: 4a5ac3c6f8383cc33c795804ba5f7f5553c029bbb4a6d28f1e4d8fb5107902c1


alt text

Figure 2: VirusTotal information.

Ransom Note

The ransom note is written in HTML format and is dropped into RecoveryManual.html files on the system.

The client ID embedded inside the ransom note is generated from the victim’s computer name and a hard-coded string in memory.

alt text

Figure 3: MountLocker ransom note.


MountLocker has pretty average performance and does not fully utitlize the machine’s processing power.

alt text

Figure 4: ANY.RUN sandbox result.

Static Code Analysis

Command Line Parameters

MountLocker can be ran with or without command line parameters. The ransomware first checks and parse the given parameters to modify its functionalities accordingly.

alt text

Figure 5: Parsing command line parameters.

Below is the list of arguments that can be supplied by the operators:

Argument Description
/LOGIN= Network username (for network encryption and worm)
/PASSWORD= Network password (for network encryption and worm)
/CONSOLE Logging through console
/NODEL No self-deletion
/NOKILL No service and process killing
/NOLOG No logging through file (this is hard-coded to be FALSE in this sample)
/SHAREALL Encrypting all shared resources (except ”\ADMIN$”)
/NETWORK Worm network type:
- w = Windows Management Instrumentation (WMI)
- s = service (requires ADMIN creds)
- others = unknown or default
/PARAMS= Command line parameters to launch executable with on other PCs (worm)
/TARGET= Path to a file or a directory to be encrypted specifically
There can be multiple target arguments
/FAST= Buffer size for fast encryption (default: 0x10000000 bytes)
/MIN= Minimum file size to encrypt (default: 0 bytes)
/MAX= Maximum file size to encrypt (default: 0 bytes)
/FULLPD Does not avoid encrypting Program Files, Program Files (x86)
ProgramData, and SQL
/MARKER= Marker file name to drop in each encrypted drive
/NOLOCK= Avoid encrypting:
- L: Local
- N: Network
- S: Network shared resources


The ransomware has two different ways to log its operations, and each can be enabled through setting the command line arguments /CONSOLE to 1 and /NOLOG to 0.

In this particular sample, /NOLOG flag’s value is hard-coded to be 0, so it always records and drops a log file on the victim’s system.

When the /NOLOG flag is 0, MountLocker extracts the current executable’s file path, append .log to the end, and use that as the log file path.

alt text

Figure 6: Creating log file in current directory.

When the /CONSOLE flag is 1, MountLocker will also log through console standard output stream. It calls AllocConsole and GetStdHandle(STD_OUTPUT_HANDLE) to allocate the console and get a handle to the standard output stream.

To write to this console, it calls WriteConsoleW with this handle.

alt text

Figure 7: Creating log file in current directory.

The beginning of the log tells us the version of the specific MountLocker sample, and in this case, the version is 5.0.

It also extracts and records information about the victim’s system such as the number of processors, total system memory, Windows version, system architecture, …

alt text

Figure 8: Logging system information.

All file and network operations (enumeration, skipping, encrypting, error) are recorded this way.

alt text

Figure 9: MountLocker log file.

Terminating Services

If the /NETWORK argument is not provided, the malware will run in local mode.

In this mode, if the /NOKILL argument is 0, it enumerates and kills all services with these strings in their name.

"SQL", "database", "msexchange"

First, it calls OpenSCManagerA to obtain a handle to the service control manager and calls EnumServicesStatusA to enumerate all Win32 services with status SERVICE_ACTIVE.

alt text

Figure 10: Enumerating through all active services.

If a service contains any of the three strings above, MountLocker will terminate it by calling OpenServiceA to obtain a service control handle and calling ControlService to send a control stop code. It then continuously loops until the service’s state is SERVICE_CONTROL_STOP to make sure the service is fully terminated.

alt text

Figure 11: Sending control stop code to terminate service.

Terminating Processes

If it’s running in local mode and the /NOKILL argument is 0, MountLocker will enumerate and kill all processes with these strings in their name.

"msftesql.exe", "sqlagent.exe", "sqlbrowser.exe", "sqlwriter.exe", "oracle.exe", "ocssd.exe", 
"dbsnmp.exe", "synctime.exe", "agntsvc.exe", "isqlplussvc.exe", "xfssvccon.exe", "sqlservr.exe", 
"mydesktopservice.exe", "ocautoupds.exe", "encsvc.exe", "firefoxconfig.exe", "tbirdconfig.exe", 
"mydesktopqos.exe", "ocomm.exe", "mysqld.exe", "mysqld-nt.exe", "mysqld-opt.exe", "dbeng50.exe", 
"sqbcoreservice.exe", "excel.exe", "infopath.exe", "msaccess.exe", "mspub.exe", "onenote.exe", 
"outlook.exe", "powerpnt.exe", "sqlservr.exe", "thebat.exe", "steam.exe", "thebat64.exe", "thunderbird.exe", 
"visio.exe", "winword.exe", "wordpad.exe", "QBW32.exe", "QBW64.exe", "ipython.exe", "wpython.exe", 
"python.exe", "dumpcap.exe", "procmon.exe", "procmon64.exe", "procexp.exe", "procexp64.exe"

The ransomware first calls ZwQuerySystemInformation with the information class of SystemProcessInformation to get an array of SYSTEM_PROCESS_INFORMATION structures. It enumerates through each running process, avoids its own process, and starts terminating processes in the kill list.

alt text

Figure 12: Enumerating through all active processes.

To check and kill a process, it loops through the PROCESS_TO_KILL list and compares the process name. If the process name is in the list, it calls OpenProcess to get the handle of that process and terminates it using TerminateProcess.

alt text

Figure 13: Terminating processes that are in the kill list.

Generating Global ChaCha20 Key

Next, it randomly generates the global ChaCha20 key. The randomization is done through calling the rdtsc instruction to get the processor time stamp and xoring its least significant byte to generate each byte in the key.

After generating the global key, the ransomware copies the key to another global buffer in memory and encrypts this new buffer using the hard-coded RSA-2048 key.

alt text

Figure 14: Randomly generate global ChaCha20 key and encrypt it with RSA-2048.

MountLocker later uses this global ChaCha20 key to encrypt and protect its ChaCha20 keys instead of using RSA-2048. Since RSA-2048 encryption is only performed once, there is some performance advantage with this hybrid-cryptography scheme since RSA is quite slow compared to ChaCha20.


Creating Encrypting Threads

Despite having different schemes for different drive types and targets, the encryption functionality is pretty much the same.

MountLocker has a specific function that takes in a drive/file name to encrypt and a function to enumerate through it as parameters.

This function first passes the enumerating function and the target name to a custom structure before spawning a thread to begin the encryption.

This thread acts as the main thread in the encryption, which recursively enumerates and provides files for children threads to encrypt.

alt text

Figure 15: Spawning main thread.

The main thread function calls CreateEventA to create an event handler for each child thread to later send them file information through calling SetEvent.

Only 2 children worker threads are spawned, and these threads loops and waits to receive files from the main thread to encrypt. The main thread will begin feeding them files by calling the enumeration function in the custom structure above and enumerating through the target folder.

alt text

Figure 16: Main thread spawning children threads and starting file enumeration.

Children Worker Threads

Once spawned, each worker thread receives a shared structure with the main thread, and it constantly loops to check for the encrypt signal is 1 in this shared structure.

Due to synchronization through sharing a common structure among threads, the child thread calls _InterlockedExchange to atomically extract the encrypt signal to check if it’s allowed to encrypt.

As it finds files to encrypt, the main thread adds the file name to the shared structure and sets the encrypt signal for the child thread to process that file.

alt text

Figure 17: Child thread waiting for encrypt signal to encrypt files.

After receiving the file information, the worker thread creates a structure to store file information such as filename, encrypted filename, file handle, file size, …

It will then checks to see if it has priviledge to open the file and retrieve the file size.

alt text

Figure 18: Checking if file can be opened.

Next, it randomly generates the file’s ChaCha20 key and appends it to the file structure above. The randomization is done through calling the rdtsc instruction similar to the global ChaCha20 key generation.

alt text

Figure 19: Randomly generating ChaCha20 key for each file.

After generating the ChaCha20 file key, the worker thread creates a 313-byte buffer that stores the file marker string “lock2” in little endian, the fast encryption size, the encrypted ChaCha20 global key, and the encrypted ChaCha20 file key. This buffer is appended at the end of the to-be-encrypted file.

alt text

Figure 20: Generating key buffer and writing it at the end of the file.

Here is the layout of the key buffer at the end of an encrypted file.

alt text

Figure 21: Key buffer layout.

File encryption is pretty standard. The worker thread encrypts a 0x100000-byte chunk at a time until it has encrypted FAST_CRYPT_SIZE bytes or ran out of bytes to encrypt.

It uses ReadFile to read file content into a buffer, encrypts it using the ChaCha20 file key, and writes it back using WriteFile. Because encryption is performed on the same file, SetFilePointerEx is called to adjust the file pointer after reading and writing.

alt text

Figure 22: ChaCha20 File Encryption.

I won’t analyze the ChaCha20 function cause MountLocker basically just uses this CRYPTOGAMS library by OpenSSL.

Main Thread Enumeration

MountLocker uses the same function for file traversal for network drives, network shares, and local drives.

Before traversing a drive, the ransomware checks if a marker file name is provided from the /MARKER= command line argument. If it is, MountLocker creates an empty file with this marker file name in the to-be-encrypted drive before enumerating it. This is mainly for marking which drive has been encrypted.

alt text

Figure 23: Creating drive marker file.

To enumerate through folders, MountLocker calls FindFirstFileW and FindNextFileW. When enumerating through network servers, it will use WNetOpenEnumW and WNetEnumResourceW instead.

alt text

Figure 24: Recursive file traversal.

The ransomware also calls a function to checks if it should encrypt each file/folder that it finds.

When processing a folder, the checking function will check for the following things. If any of these is true, the folder is skipped.

  - If folder name is "." or ".."
  - If folder name is in the FOLDER_TO_AVOID list
  - If folder name is "Program Files", "Program Files (x86)", "ProgramData", or "SQL"
  - If calling CreateFileW on the folder fails.
  - If folder's reparse tag is not IO_REPARSE_TAG_MOUNT_POINT (folder is a mount point) 
  or IO_REPARSE_TAG_SYMLINK (folder is a symbolic link)\
  - If folder name is in a share name format
  - If folder is a mount point and is visible

Below is the FOLDER_TO_AVOID list.

":\\Windows\\", ":\\System Volume Information\\", ":\\$RECYCLE.BIN\\", ":\\SYSTEM.SAV", ":\\WINNT", 
":\\$WINDOWS.~BT\\", ":\\Windows.old\\", ":\\PerfLog\\", ":\\Boot", ":\\ProgramData\\Microsoft\\", 
":\\ProgramData\\Packages\\", "$\\Windows\\", "$\\System Volume Information\\", "$\\$RECYCLE.BIN\\", 
"$\\SYSTEM.SAV", "$\\WINNT", "$\\$WINDOWS.~BT\\", "$\\Windows.old\\", "$\\PerfLog\\", "$\\Boot", 
"$\\ProgramData\\Microsoft\\", "$\\ProgramData\\Packages\\", "\\WindowsApps\\", "\\Microsoft\\Windows\\", 
"\\Local\\Packages\\", "\\Windows Defender", "\\microsoft shared\\", "\\Google\\Chrome\\", "\\Mozilla Firefox\\", 
"\\Mozilla\\Firefox\\", "\\Internet Explorer\\", "\\MicrosoftEdge\\", "\\Tor Browser\\", "\\AppData\\Local\\Temp\\"

If the folder is valid and there is no ransom note file in the folder yet, MountLocker will drop a ransom note in the folder.

alt text

Figure 25: Dropping ransom note.

When processing a file, the checking function checks for the following things. If any of these is true, the file is skipped.

  - If file size is less than MIN_CRYPT_SIZE (if MIN_CRYPT_SIZE is provided)
  or if file size is larger than MAX_CRYPT_SIZE (if MAX_CRYPT_SIZE is provided)
  - If file name is "RecoveryManual.html", "bootmgr", or has the encrypted file extension.
  - If file extension is in the EXTENSION_TO_AVOID list

Below is the EXTENSION_TO_AVOID list.

"exe", "dll", "sys", "msi", "mui", "inf", "cat", "bat", "cmd", "ps1", "vbs", "ttf", "fon", "lnk"

If the file is valid, the ransomware’s main thread will populate the shared file structure with the file name for its worker thread to encrypt.

Because of synchronization concerns, the main thread also has to call WaitForSingleObject and _InterlockedExchange to wait until it has access to the shared structure.

After populating the file structure, it calls SetEvent to signal the event for worker threads to encrypt.

alt text

Figure 26: Calling SetEvent to signal file encryption.

Worm Property

Similar to WannaCry and Ryuk, this MountLocker sample is a combination of ransomware and worm with the ability to self-propagate to other hosts in the network.

Unlike WannaCry, this ransomware does not use any fancy 0-day but instead just COM interfaces such as IDirectorySearch and IWbemServices to spread and execute itself.

MountLocker has this structure that is shared among all worm threads.

  _QWORD function; // function to launch ransomware remotely
  _QWORD func_param; // function's parameter
  HANDLE hEvent; // worm event
  HANDLE hSemaphore; // worm semaphore

First, memory is allocated for this structure, and the event handle and semaphore handle are created. The ransomware launching function and its parameter is originally left to be null initially.

MountLocker creates 8 threads to execute this worm property.

alt text

Figure 27: Populating worm struct and creating worm threads.

Each of these threads waits for the event to be signal by the main thread before calling the worm function to execute the ransomware remotely. The main thread will set this worm function accordingly before signalling the event.

alt text

Figure 28: Worm worker threads.

After creating these worker threads, the main thread begins enumerating the Windows domain that the current host is in.

This is accomplished through calling NetGetDCName to get the name of the primary domain controller and append this name after the string “LDAP://”.

alt text

Figure 29: Building LDAP path.

Lightweight Directory Access Protocol (LDAP) is a protocol to communicate and query several different types of directories, and in this case, MountLocker uses it to make Active Directory query requests to the primary domain controller.

It calls ADsOpenObject with the newly built ADsPath string and provides the credential (username and password) from the /LOGIN= and /PASSWORD= arguments. The RIID provided is {109BA8EC-92F0-11D0-A790-00C04FD8D5A8}, and through this call, the ransomware retrieves the IDirectorySearch interface.

This trick to query IDirectorySearch is previously used by Trickbot as explained by Vitali here.

alt text

Figure 30: Querying IDirectorySearch interface.

This interface can be used to execute a search for all domain controllers through its IDirectorySearch::ExecuteSearch function which return an ADs search handle.

MountLocker calls IDirectorySearch::GetFirstRow and IDirectorySearch::GetNextRow to enumerate through all the searches, passing each search into a function to extract its domain controller information.

alt text

Figure 31: Enumerating through ADs searches to extract domain controller information.

For each of these search handles, MountLocker then calls IDirectorySearch::GetColumn with the column name “name” to retrieve the corresponding ADS_SEARCH_COLUMN structure at this row.

This structure contains an array of ADSVALUE structures, and each of these structures contains a DN string of a directory service object in the Active Directory. This Distinguished Name (DN) string is basically a name to identify another PC in the network.

alt text

Figure 32: Extracting all DN string of other PCs in the network.

When a DN string of a PC is extracted, it’s passed into a function where the ransomware will use it as the function parameter in the WORM_STRUCT structure. The structure’s function is set to a specific function that drops and launches the sample remotely. SetEvent is called to execute this function after the WORM_STRUCT structure is fully populated.

alt text

Figure 33: Setting up WORM_STRUCT and signal the worm event.

Worm Dropping Function

First, the worm thread will try to establish a connection to the remote target PC by calling WNetAddConnection2W and provice the username and password from the /LOGIN= and /PASSWORD= arguments.

alt text

Figure 34: Establishing connection with remote PC.

Next, memory is allocated for a custom structure. I just call this WORM_REMOTE_STRUCT.

  LPCWSTR rem_exe_path; // remote executable path
  CHAR *launch_exe_cmd; // command line to launch executable
  CHAR *PC_name; // remote PC name
  CHAR *elevated_PC_path; // Elevated PC path to launch executable
  DWORD API_result; // result value
  DWORD last_error; // last error value
  CHAR *exe_name; // executable name

It then populates this structure. The executable name is a number retrieved from GetTickCount, and the path on the host to drop the ransomware is set to “C:\ProgramData”.

alt text

Figure 35: Populating WORM_REMOTE_STRUCT.

The drop_ransomware function checks if the DN string contains either of the share names with higher priviledge ”\ADMIN$“ and ”\IPC$“. If it does, then MountLocker uses that as the main path in the command to launch the executable. If it doesn’t, then it just uses the normal path.

The ransomware sample is set to be launched with the /NOLOG parameter and any arguments provided in the original /PARAMS= argument.

Finally, it drops the ransomware on the target PC by calling CopyFileW.

alt text

Figure 36: Dropping the ransomware on the target PC.

Not only does MountLocker drops the ransomware executable on the target PC but it also enumerates through the PC’s shared resources in the PC’s network by calling NetShareEnum. After finding the path to each shared resource, the ransomware calls drop_ransomware to drop the executable in the shared resource’s system.

alt text

Figure 37: Dropping the ransomware on the target PC’s shared resources.

Worm Launching Function

MountLocker has two different ways to launch the executable on the remote host.

If the /NETWORK argument provided is s, it launches the executable through a service.

First, this full cmd.exe command is built.

cmd.exe /c start "ransomware_path PARAMS_VALUE /NOLOG"

Then, the ransomware calls OpenSCManagerW to establish a connection to the service control manager on the target PC. Using this handle, it calls CreateServiceW with the command above as its lpBinaryPathName parameter to create a service handle and calls StartServiceW to launch it.

alt text

Figure 38: Launching ransomware on remote host using Service.

If the /NETWORK argument provided is w, it launches the executable through Windows Management Instrumentation (WMI).

First, MountLocker retrieves the IWbemServices interface. This is done by calling CoCreateInstance with the CLSID {4590F811-1D3A-11D0-891F-00AA004B2E24} to retrieve an IWbemLocator object.

Using this IWbemLocator object, it calls the IWbemLocator::ConnectServer to connect with the PC’s ROOT\CIMV2 namespace and obtain an IWbemServices object.

alt text

Figure 39: Connecting to ROOT\CIMV2 namespace through COM objects.

From here, MountLocker sets up an appropriate SEC_WINNT_AUTH_IDENTITY_A structure with the given username and password. It then calls CoSetProxyBlanket to set the authentication information for this IWbemServices object.

alt text

Figure 40: Setting the authentication information for the IWbemServices object.

Using this IWbemServices object, the ransomware calls the IWbemServices::GetObjectA function with the “Win32_Process” path to get IWbemClassObject object corresponding to Windows32 processes.

Next, using this “Win32_Process” object, it then calls the IWbemClassObject::GetMethod function with the “Create” method name to get an IWbemClassObject object corresponding to the method to create a process.

With this method object, it calls the IWbemClassObject::SpawnInstance to create a new instance of the class.

alt text

Figure 41: Retrieving the COM object to create a Windows32 process.

Since the Win32_Process::Create requires a valid value for the command line in-parameter to execute properly, MountLocker calls the IWbemClassObject::Put function to set the value of the command line to the launching command that it has built above.

alt text

Figure 42: Setting valid value for command line in-parameter.

Finally, it calls IWbemServices::ExecMethod to create a Win32 process running the “cmd.exe” command above. It also checks to see if the new process is created successfully or not by checking if the process’s ID is changed through calling IWbemClassObject::Get.

alt text

Figure 43: Launching ransomware remotely using Win32_Process::Create.

If any of these steps to drop and launch the executable fails, MountLocker just resorts to using WNetOpenEnumW and WNetEnumResourceW to enumerate through the victim’s network and drops the ransomware in a similar fashion.


If the /NODEL argument is set to 0, MountLocker will delete its own executable.

First, it creates a .bat file in the TEMP folder with a random name from GetTickCount.

It writes this command into this .bat file, which clears Read-only, System, and Hidden file attribute from the ransomware executable, forces deletes the executable quietly if it exists, and deletes the bat file.

attrib -s -r -h %1
del /F /Q %1
if exist %1 goto l
del %0

Next, MountLocker builds the command line string to execute the .bat file with the executable path as the parameter and finally calls CreateProcessW to delete itself.

alt text

Figure 44: Self-deletion.

YARA rule

rule MountLocker5_0 {
		description = "YARA rule for MountLocker v5.0"
		reference = ""
		author = "@cPeterr"
		tlp = "white"
		$worm_str = "========== WORM ==========" wide
		$ransom_note_str = ".ReadManual.%0.8X" wide
		$version_str = "5.0" wide
		$chacha_str = "ChaCha20 for x86_64, CRYPTOGAMS by <>"
		$chacha_const = "expand 32-byte k"
		$lock_str = "[OK] locker.file > time=%0.3f size=%0.3f KB speed=%" wide
		$bat_str = "attrib -s -r -h %1"
		$IDirectorySearch_RIID = { EC A8 9B 10 F0 92 D0 11 A7 90 00 C0 4F D8 D5 A8 }
		uint16(0) == 0x5a4d and all of them