Develop Reference

Disable specific USB Storage Device C#

I am working on a project that needs to enable/disable specific USB storage devices.
My app gets all storage USB DeviceIDs, and depending on saved settings then needs to allow the device or not. I have used the code from this previous question for the enable/disable, and call it like so:
DisableHardware.DisableDevice(n => n.ToUpperInvariant().Contains(VidPid), true);
with string VidPid = "VID_8564&PID_1000";.
I have stepped through the code, it all works perfectly (as far as I can tell) until the SetupDIChangeState call, which then returns -536870389 (E000020B as hex) as error code (from Marshall.GetLastWin32Error()).
Apparently this refers to either 1) the device not being present (which as far as I understand is not the case here, as all other calls in this class work fine, and I get `VidPid' from
private static List<WMUBClasses.USBDeviceInfo> GetUSBDevices()
{
try
{
List<WMUBClasses.USBDeviceInfo> tList = new List<WMUBClasses.USBDeviceInfo>();
ManagementObjectCollection collection;
using (var searcher = new ManagementObjectSearcher(#"SELECT * FROM Win32_USBHub"))
{
collection = searcher.Get();
}
foreach (var device in collection)
{
if (!device.GetPropertyValue("Description").ToString().Contains("Storage"))
{
continue;
}
tList.Add(new WMUBClasses.USBDeviceInfo(
(string)device.GetPropertyValue("DeviceID"),
(string)device.GetPropertyValue("PNPDeviceID"),
(string)device.GetPropertyValue("Description")
));
}
return tList;
}
catch (Exception ex)
{
return null;
}
}
or 2) an incorrect build platform, but I have tried all the different combinations (Any CPU, Mixed Platforms, x86 and x64), they all return the same result.
I have also looked at this which is another approach to my problem (by creating and using a kernel mode filter driver), it just seems like killing a fly with a wrecking ball. To be honest I have no clue of how to go about using this (for someone that hasn't done any driver development it looks super intimidating, especially after having read some of the available documentation.)
Should I (A) keep using the SetupDi API calls to achieve my goal and if so, can anyone see what is wrong with the code or how I am using it? If not (A), should I (B) use the filter driver approach instead and if so, any pointers?
As stated in the header, I want to disable specific USB storage devices, so as far as I understand, this precludes using the Registry to disable ALL USB storage devices. So if neither of the above, does anyone have any other direction I should be looking at instead?

Is it possible to get PC's # of physical USB ports?

I'm trying to get number of physical USB ports of PC for different Windows operating systems. To start with it seemed like an easy task but now I'm getting desperate.
Maybe this question is not even valid since I don't know how USB ports are treated on hardware level.
I thought using WMI(C) and ManagementObjectSearcher would be the right path to take and it returned right values... on certain operating systems. Or that's what I thought.
For example, I tried the following:
// >wmic path win32_usbhub get name
private const string UsbProperty = "Name";
private const string UsbPath = "Win32_USBHub";
private const string UsbPortName = "USB ROOT HUB";
...
// Get USB Ports
public IEnumerable<string> GetUsbPorts()
{
// All from Win32_USBHub where name contains USB ROOT HUB
return GetManagementObjectValues(UsbProperty, UsbPath).Where(name =>
CultureInfo.CurrentCulture.CompareInfo.IndexOf(
name, UsbPortName, CompareOptions.IgnoreCase) >= 0);
}
.
// Query
private static IEnumerable<string> GetManagementObjectValues(
string properties, string path, Func<object, string> formatter = null)
{
var values = new List<string>();
string query = string.Format("SELECT {0} FROM {1}", properties, path);
var search = new ManagementObjectSearcher(query);
try
{
foreach (ManagementObject item in search.Get())
{
string value = string.Empty;
foreach (string property in properties.Split(',')
.Select(prop => prop.Trim()))
{
if (item[property] == null)
continue;
if (value.Length > 0)
value += " ";
if (formatter != null)
value += formatter(item[properties]);
value += item[property].ToString();
}
values.Add(value.TrimEnd());
}
}
catch (Exception e)
{
if (e is ManagementException)
Logger.Warn(string.Format(
"Can't extract {0} properties of {1}", properties, path), e);
else
Logger.Error(e);
}
return values.Count >= 1 ? values : new List<string> { DefaultValue };
}
This seemed to get me the right amount on Windows8 but on WindowsXP it was totally off.
Next, I tried (for example) the following. I noticed that on Win8 I have USB<number> as ExternalReferenceDesignator but on WinXP, there's plain USB as InternalReferenceDesignator and external is empty.
For XP this seemed to work just fine but then again on Win8 amount of ports was six (6). Actual port count is 3 and with docking station station seven (7).
// >wmic path Win32_PortConnector get ExternalReferenceDesignator,InternalReferenceDesignator
private const string UsbPortName = "USB";
private const string PortProperties =
"ExternalReferenceDesignator, InternalReferenceDesignator";
private const string PortPath = #"Win32_PortConnector";
...
public IEnumerable<string> GetEthernetPorts()
{
// All where external includes USB or internal equals USB
return GetManagementObjectValues(PortProperties, PortPath).Where(name =>
CultureInfo.CurrentCulture.CompareInfo.IndexOf(
name, UsbPortName, CompareOptions.IgnoreCase) >= 0 ||
string.Compare(name, UsbPortName, StringComparison.OrdinalIgnoreCase) == 0);
}
So is it even possible or am I just looking from the wrong place?

And to answer my own question: No, it's not possible.
WMIC query results for Win32_USBController (or some related path) might seem right but you can't draw any conclusions from them. Information about connectors aren't stored on the baseboard or any other location either.
For example an old Dell Latitude D830 with Windows XP SP3 has three (3) physical connectors but WMIC and USBView.exe shows results below:
C:\>wmic path win32_usbcontroller get caption
Caption
Intel(R) ICH8 Family USB Universal Host Controller - 2834
Intel(R) ICH8 Family USB Universal Host Controller - 2835
Intel(R) ICH8 Family USB2 Enhanced Host Controller - 283A
Intel(R) ICH8 Family USB Universal Host Controller - 2830
Intel(R) ICH8 Family USB Universal Host Controller - 2831
Intel(R) ICH8 Family USB Universal Host Controller - 2832
Intel(R) ICH8 Family USB2 Enhanced Host Controller - 2836

I know it has been a while since the question was asked, but I just happened to be working on a USB port tree for a support app of some sort.
Initially, I tried to use Win32..., as much as it is not intended to be used as a source of information for the rebuilding of the device tree, the answer of this post explains some possibilities (Get List of connected USB Devices).
I played with it but I did not like it. UsbView by itself was also not an option (lack of skill in C ). Luckily I found NativeUsbLib. It provides you with a USB device tree, you just need to understand how to read it. Not a physical USB port This image shows a port that clearly states it is not possible to use it. In addition to that parameter, there is "Companion Hub Symbolic Link Name", in my experience, it has to be present on a valid USB port.
As for multiple controllers, and even multiple ports in one controller that satisfy my previous statement, there is an explanation. Multiple ports for same physical port, in my case, ports 1 and 13 are the same. If a device is 1.1 or 2.0 it will show under port 1, and if it supports 3.0 it will show up under port 13. And same goes for two controllers, they don't have 100% the same structure, but once you strip unnecessary data, and merge data (not necessarily all of it) that is left, the result will be a USB port tree. I can't guarantee all of the statements are true, but that is my current experience, and I might update this post. If not, feel free to message me if you have some questions. Btw. NativeUsbLib by default does not provide data for ports that don't have a device present (plugged in). To fix that, comment out lines in DeviceFactory that check if the device is present (line 35).
Hope this helps someone and sorry for my spelling, I am sure I messed up somewhere...

Retrieve host VM MAC address via guest VM

This has been an issue I have been looking to for two days. I will share my findings.
I am currently working on an in-house license management system for our software. It's nothing too fancy - as long as it can uniquely identify a user, it's good enough. Our mechanism currently relies on user sign-in + password + MAC address.
99% of the users so far have had no issues, but there is a small subset, the 1%, that has been returning an issue. This 1% is so important to us, because one failure means one hole in our system, something we would like to weed out. Okay - onto the main topic.
Method 1:
public static string returnMAC1()
{
ManagementObjectSearcher searcher = new ManagementObjectSearcher("Select MACAddress, PNPDeviceID FROM Win32_NetworkAdapter WHERE MACAddress IS NOT NULL AND PNPDEVICEID IS NOT NULL");
ManagementObjectCollection mObject = searcher.Get();
foreach (ManagementObject obj in mObject)
{
string pnp = obj["PNPDeviceID"].ToString();
if (pnp.Contains("PCI\\"))
{
string mac = obj["MACAddress"].ToString();
mac = mac.Replace(":", string.Empty);
return mac;
}
}
return "Nothing happened...";
}
Method 1 retrieves the MAC address based on the fact that the physical card is connected to the PCI interface.
Method 2:
public static string returnMAC2()
{
string mac = string.Empty;
foreach (System.Net.NetworkInformation.NetworkInterface nic in System.Net.NetworkInformation.NetworkInterface.GetAllNetworkInterfaces())
{
if (nic.OperationalStatus == System.Net.NetworkInformation.OperationalStatus.Up)
{
macAddresses += nic.GetPhysicalAddress().ToString();
break;
}
}
return mac;
}
The second method is a standard method retrieved from MSDN documentation in-regards to MAC addresses.
Based on some tests, it seems the second method is a tad unreliable to retrieve MAC addresses, since it retrieves the wireless card's address. We've had some users returning null addresses as a result of using that method, and while I don't know why that would happen, it could be because there's a lack of a wireless card in their computers. With that said, that's only conjecture. Method #1 relies on using SQL queries to retrieve the PCI MAC. This one has been reliable.
Tests:
Using a Windows 8.1 Enterprise Evaluation edition (free 90-day trial, yay!) installed onto the VirtualBox VM, the tests confirmed that there are major differences in the MAC addresses returned via the guest VM and the host VM.
According to my research, in most cases, the virtual machine is assigned the same MAC address every time it is powered on, so long as the virtual machine is not moved and no changes are made to the certain settings in the configuration file. With that said, and here's the bad news... The guest VM MAC could be anything. So it seems, this is one of the reasons the MAC addresses are inconsistent when used as unique identifiers, which is an issue I found out when some users were on their company VMs. I never knew that's the way people worked, but here we are, so no gloating about it at this point.
My question is - is there any way, without making the user change any settings on their end, to retrieve the host VM's MAC as opposed to the guest VM?
At this point I don't see any reason why someone won't assign the same MAC to every single guest machine to get around our floating license mechanism. Retrieving the host VM MAC would get around this issue, as it would show as one MAC.

We decided this is both impossible and unnecessary. We also decided to use the motherboard UUID as the primary unique identifier, and falling back to the MAC address using the MAC address method below, in case the UUID returns a FFFF-FFFF....... on the rare occasion the vendor does not supply a UUID to that motherboard.
public static string returnMAC1()
{
ManagementObjectSearcher searcher = new ManagementObjectSearcher("Select MACAddress, PNPDeviceID FROM Win32_NetworkAdapter WHERE MACAddress IS NOT NULL AND PNPDEVICEID IS NOT NULL");
ManagementObjectCollection mObject = searcher.Get();
foreach (ManagementObject obj in mObject)
{
string pnp = obj["PNPDeviceID"].ToString();
if (pnp.Contains("PCI\\"))
{
string mac = obj["MACAddress"].ToString();
mac = mac.Replace(":", string.Empty);
return mac;
}
}
return "Nothing happened...";
}

Is there really any way to uniquely identify any computer at all

I know there are a number of similar questions in stackoverflow such as the followings:
What's a good way to uniquely identify a computer?
What is a good unique PC identifier?
Unique computer id C#
WIN32_Processor::Is ProcessorId Unique for all computers
How to uniquely identify computer using C#?
... and dozens more and I have studied them all.
The problem is that some of the accepted answers have suggested MAC address as an unique identifier which is entirely incorrect. Some other answers have suggested to use a combination of various components which seems more logical. However, in case of using a combination it should be considered which component is naturally unlikely to be changed frequently. A few days ago we developed a key generator for a software licensing issue where we used the combination of CPUID and MAC to identify a windows pc uniquely and till practical testing we thought our approach was good enough. Ironically when we went testing it we found three computers returning the same id with our key generator!
So, is there really any way to uniquely identify any computer at all? Right now we just need to make our key generator to work on windows pc. Some way (if possible at all) using c# would be great as our system is developed on .net.
Update:
Sorry for creating some confusions and an apparently false alarm. We found out some incorrectness in our method of retrieving HW info. Primarily I thought of deleting this question as now my own confusion has gone and I do believe that a combination of two or more components is good enough to identify a computer. However, then I decided to keep it because I think I should clarify what was causing the problem as the same thing might hurt some other guy in future.
This is what we were doing (excluding other codes):
We were using a getManagementInfo function to retrieve MAC and Processor ID
private String getManagementInfo(String StrKey_String, String strIndex)
{
String strHwInfo = null;
try
{
ManagementObjectSearcher searcher = new ManagementObjectSearcher("select * from " + StrKey_String);
foreach (ManagementObject share in searcher.Get())
{
strHwInfo += share[strIndex];
}
}
catch (Exception ex)
{
// show some error message
}
return strHwInfo;
}
Then where needed we used that function to retrieve MAC Address
string strMAC = getManagementInfo("Win32_NetworkAdapterConfiguration", "MacAddress");
and to retrieve ProcessorID
string strProcessorId = getManagementInfo("Win32_Processor", "ProcessorId");
At this point, strMAC would contain more than one MAC address if there are more than one. To take only one we just took the first 17 characters (12 MAC digits and 5 colons in between).
strMAC = strMAC.Length > 17 ? strMAC.Remove(17) : strMAC;
This is where we made the mistake. Because getManagementInfo("Win32_NetworkAdapterConfiguration", "MacAddress") was returning a number of extra MAC addresses that were really in use. For example, when we searched for MAC addresses in the command prompt by getmac command then it showed one or two MAC addresses for each pc which were all different. But getManagementInfo("Win32_NetworkAdapterConfiguration", "MacAddress") returned four to five MAC addresses some of which were identical for all computers. As we just took the first MAC address that our function returned instead of checking anything else, the identical MAC addresses were taken in strMAC incidently.
The following code by Sowkot Osman does the trick by returning only the first active/ enabled MAC address:
private static string macId()
{
return identifier("Win32_NetworkAdapterConfiguration", "MACAddress", "IPEnabled");
}
private static string identifier(string wmiClass, string wmiProperty, string wmiMustBeTrue)
{
string result = "";
System.Management.ManagementClass mc = new System.Management.ManagementClass(wmiClass);
System.Management.ManagementObjectCollection moc = mc.GetInstances();
foreach (System.Management.ManagementObject mo in moc)
{
if (mo[wmiMustBeTrue].ToString() == "True")
{
//Only get the first one
if (result == "")
{
try
{
result = mo[wmiProperty].ToString();
break;
}
catch
{
}
}
}
}
return result;
}
//Return a hardware identifier
private static string identifier(string wmiClass, string wmiProperty)
{
string result = "";
System.Management.ManagementClass mc = new System.Management.ManagementClass(wmiClass);
System.Management.ManagementObjectCollection moc = mc.GetInstances();
foreach (System.Management.ManagementObject mo in moc)
{
//Only get the first one
if (result == "")
{
try
{
result = mo[wmiProperty].ToString();
break;
}
catch
{
}
}
}
return result;
}
However, I was absolutely right about the identical Processor ID issue. All three returned the same Processor ID when we put wmic cpu get ProcessorId command in their command prompts.
Now we have decided to use Motherboard serial number instead of Processor ID to make a combination with MAC address. I think our purpose will be served with this way and if it doesn't in some cases then we should let it go in those few cases.

How about adding motherboard serial number as well e.g.:
using System.management;
//Code for retrieving motherboard's serial number
ManagementObjectSearcher MOS = new ManagementObjectSearcher("Select * From Win32_BaseBoard");
foreach (ManagementObject getserial in MOS.Get())
{
textBox1.Text = getserial["SerialNumber"].ToString();
}
//Code for retrieving Processor's Identity
MOS = new ManagementObjectSearcher("Select * From Win32_processor");
foreach (ManagementObject getPID in MOS.Get())
{
textBox2.Text = getPID["ProcessorID"].ToString();
}
//Code for retrieving Network Adapter Configuration
MOS = new ManagementObjectSearcher("Select * From Win32_NetworkAdapterConfiguration");
foreach (ManagementObject mac in MOS.Get())
{
textBox3.Text = mac["MACAddress"].ToString();
}

The fact in getting a globally unique ID is, only MAC address is the ID that will not change if you set up your system all over. IF you are generating a key for a specific product, the best way to do it is assigning unique IDs for products and combining the product ID with MAC address. Hope it helps.

I Completely agree with just the above comment.
For Software licensening, you can use:
Computer MAC Address (Take all if multiple NIC Card) + Your software Product Code
Most of the renowned telecom vendor is using this technique.

However, I was absolutely right about the identical Processor ID
issue. All three returned the same Processor ID when we put wmic cpu
get ProcessorId command in their command prompts.
Processor ID will be same if all the systems are running as virtual machines on the same hypervisor.
MAC ID seems fine. Only thing is users must be provided the option to reset the application, in case the MAC changes.

It looks like custom kitchen is the way for that.
SMBIOS UUID (motherboard serial) is not robust, but works fine in 99% cases. However some brands will set the same UUID for multiple computers (same production batch maybe). Getting it requires WMI access for the user (if he's not administrator), you can solve that by starting an external process asking administrator priviledges (check codeproject.com/Articles/15848/WMI-Namespace-Security)
Windows Product ID might be good, but I read it could be identical in some circumstances (https://www.nextofwindows.com/the-best-way-to-uniquely-identify-a-windows-machine)
Could someone clarify if the same Product ID (not product key) might be present on multiple computers ?
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Cryptography\MachineGuid seems interesting. It's generated when installing Windows and if changed, it requires to reactivate Windows.
Mac Addresses are interresting but you can only take the first one or your unique ID will change when the interface is disabled, or when another network interface is added and appears first etc.
Hard Drive serial number is nice but when installing a ghost, it might also override the serial number from the original drive... And the HD serial is very easy to change.
The best might be to generate an ID with a combination of those machine identifiers and decide if the machine is the same by comparing those identifiers (ie if at least one Mac address + either SMBIOS UUID or Product ID is ok, accept)

Unique computer ID

I'm looking for a way to get unique computer ID.
According to this post I can't use processor ID for this purpose. Can I take motherboard ID? What is the best way to identify the computer?

Like you've said CPU Id wont be unique, however you can use it with another hardware identifier to create your own unique key.
Reference assembly System.Management
So, use this code to get the CPU ID:
string cpuInfo = string.Empty;
ManagementClass mc = new ManagementClass("win32_processor");
ManagementObjectCollection moc = mc.GetInstances();
foreach (ManagementObject mo in moc)
{
cpuInfo = mo.Properties["processorID"].Value.ToString();
break;
}
Then use this code to get the HD ID:
string drive = "C";
ManagementObject dsk = new ManagementObject(
#"win32_logicaldisk.deviceid=""" + drive + #":""");
dsk.Get();
string volumeSerial = dsk["VolumeSerialNumber"].ToString();
Then, you can just combine these two serials to get a uniqueId for that machine:
string uniqueId = cpuInfo + volumeSerial;
Obviously, the more hardware components you get the IDs of, the greater the uniqueness becomes. However, the chances of the same machine having an identical CPU serial and Hard disk serial are already slim to none.

MAC address of the network adapter?
Security identifier (SID) of the windows OS install? (assuming it's windows you're dealing with)
Could you just generate a GUID for each PC?
What exactly are you trying to achieve?

The motherboard ID is a pretty unique identifier. Another option is to use the network cards MAC address, which are pretty much unique.