I am posting here as this is primarily a Security/Permissions in .Net question, and no other forum seems to fit.

I want to use the CustomACE class to create an ACE for the MANDATORY_INTEGRITY_LOW permission, I figure I can use the CustomACE class for this, but I have no idea how to initialize it.  A working sample would be helpful, or any other explanation on how to use this.

My intent is to create a "helper" executable for a IE add-on.  With Vista/Win7, add-ons now run in a low integrity state, and cannot normally communicate with various resources.  However a helper executable can be launched.  My intent is then to create a Remoting channel using the IpcChannel so that they can communicate.  However the IpcChannel needs to be opened with the MANDATORY_INTEGRITY_LOW ACE so that the add-ons can open the channel.

Because the MANDATORY_INTEGRITY_LOW permission is fairly new (Vista and onwards), it doesn't seem to have made it into the .Net framework yet, so the only option thatI can see is to use the CustomACE class to "roll my own".

Any help would be most appreciated.

Hi,

Can anyone tell me how to create a publisher policy file? I am having an oracle.dataaccess dll from oracle. but i want to redirect it to latest version. I dont know where i have to create and save publisher.config file and how to take a dll from it? even i dont know how to prepare keypair vale for it?

Could you please tell me stap by step?

Regards

Chellu.

Hi,

Is there a prescribed approach for calling .NET Framework 4.0 code from existing .NET 2.0 code directly? 

What are the Pros and Cons of this? 

Regards,

Vikram

Blog: http://dotnetupdate.blogspot.com |

There is a bug mentioned in David Broman's response to this thread:
http://social.msdn.microsoft.com/Forums/vstudio/en-US/c1c37c3d-d059-4328-82fd-bef1758e2c94/asynchronous-walking-of-the-net-managed-stack-using-dostacksnapshot-sampling-profiler-does-not

Can someone say exactly what version of the CLR this was fixed in?  In that thread he says it was "fixed in the CLR 4.5 beta" but if I understand CLR versioning correctly (I may not) it never got to 4.5 but instead stayed on 4.0.30319.* with patch fixes after that.  I want to be able to identify if our customers are running into this issue or something else but to do that I need to know what to look for to identify whether or not they have the fix referred to in the above post.

I believe the .NET 4.5 Beta (Consumer Preview) was CLR version 4.0.30319.17379, so perhaps it was fixed in that version and up but I want to make sure before I commit to asking my users experiencing problems to update.

Hello,everyone.

(My english is bad)

Anyone can tell me how to solve this problem?My website throw out this exception almost everyday,and then responses nothing.I have to restart the server.I use linux+jexus.How to solve this problem?

Thank you very much!

I am trying to activate a COM-visible .NET assembly using the activation context API from SilverLight using the following configuration:

IE ->

SilverLIght component ->
AssemblyLoader.dll -
                unmanaged (not-COM)
                [manifest with dependency on AssemblyDLL's manifest] ->                                                                                                                                                                                              AssemblyNET.dll (COM visible)[manifest]

The manifests are not embedded. The activation context succeeds by loading the AssemblyLoader.dll.manifest but the result from CoCreateInstance of any object from AssemblyNET.dll is always 

I can successfully call the CoCreateInstance in the following configuration:

Test.exe using the AssemblyNET.dll.manifest to create the activation context -> AssemblyNET.dll (COM visible)[manifest]

Is this kind of activation context setup possible?

Hi folks,

I'm using perfview for .net memory leak analysis and I was wondering what unreachable memory is representing?  I see a sizable difference in unreachable memory between two dumps that is considerable so I was curious.  Would this be the same as "private bytes" using other tools like vmmap from sysinternals?

thanks,

JJ

I use C# in VS 2008 framework 3.5 and VS2005 framework 2.0.

When I use system timer, I find in every several seconds it increases 8K memory usage and keep increasing as long as you wish. Here is the code. You can cut and paste it in console application in either vs2008 or vs2005, compile and run it. Then you can observe the memory increase. Does anyone have a solution or work around?

Using System;

using System.Timers;

publicclassTimer1

{

publicstaticvoid Main()

{

System.Timers.Timer aTimer = new System.Timers.Timer();

aTimer.Elapsed += newElapsedEventHandler(OnTimedEvent);

aTimer.Interval =75;

aTimer.Enabled = true;

Console.WriteLine("Press the Enter key to exit the program.");

Console.ReadLine();

GC.KeepAlive(aTimer);

}

privatestaticvoid OnTimedEvent(object source, ElapsedEventArgs e)

{

Console.WriteLine("Hello World!");

}

}

Hi All,

I am working on assignment for .NET project requiring Obfuscation, here we are working on specific modules comprising 12 dll's(all of .NET dll's), I need clarification on few aspects related to .NET Obfuscation of Dll's (since I am new to providing obfuscation for .NET dll's).

Using Dotfuscator tool (professional editon 4.5v) in GUI Standalone mode,

1. How we need to ensure obfuscation of dll in release mode., since, the dll obfuscation to be effected in Debug mode, could not the same dll's hold good in release mode too ?

I heard the Debug mode would be discarded by the Obfuscation tool(DotFuscator tool).

2. If multiple binaries are output by the build for same assembly/module.,

assuming C:\MyProject\Updatedll\Bin\UpdateDll1.dll   and similarly C:\MyProject\SubProject\MinorUpdate\UpdateDll1.dll  both these dll's are one and same, but when I build the project using DotObfuscator, I need to run the tasks again for above dll's(UpdateDll1.dll) if could it be uniquely handled ?

3. What are procedures/steps to load/build assemblies of dll's only instead of exe's in DotObfuscator tool.

Details: VS 2010 IDE, Windows 7 O/S. Using Perforce Version control tool.

Any help in resolving the above would be much appreciable.

Thanks & Regards,

VishalK_99

Vishal_K89

Hi,

I am using the forum's recommended way (http://social.msdn.microsoft.com/forums/en-US/csharpgeneral/thread/24792cdc-2d8e-454b-9c68-31a19892ca53/) of detecting a 32-bit program running under Wow64 on a 64-bit Vista system, and it used to work - it reported a 32-bit process running under Wow64 on a 64-bit system.  After a couple of months (I don't know exaclty when), this method is now reporting that - when I compile my class as x86 (32-bit) in Visual Studio 2010 SP1 - the method IsWow64Process does not exist in Kernel32.dll.  I am running Windows Vista Ultimate SP2 on a 64-bit HP PC.  I first check if the IntPtr is 8 bits - which it isn't because I have a 32-bit process.  Then I check if the Kernel32.dll is present - it is.  Then I check if the IsWow64Process method is present in Kernetl32.dll - and it is NOT.  I am positive this used to work a couple of months ago.  I moved the methods to a separate class (NativeMethods) to satisfy the Code Analysis tool, but no other changes were made.  If I compile the classe as x64, then the same code reports a 64-bit process because the IntPtr is 8 bits big.

Code extract:

{

 // The Kernelt32.dll module which can contain the "IsWow64Process" method exists.  Now see if the method "IsWow64Process" exists in this module.
                    IntPtr methodHandle = GetProcAddress(moduleHandle, "IsWow64Process");
                    if (methodHandle == IntPtr.Zero)
                    {
                        // The method "IsWow64Process" does not exist, so we are running on a 32-bit system.  This is NOT true as I am running on a 64-bit machine.
                        // Set the system architecture.
                        systemArch = "32-bit";
                    }
                    else
                    { // code which should be reached but is not... }
}

Can the way Visual Studio is told to compile (32 or 64 bit, AnyCPU, Target Platform) affect the way this works ?

Any help appreciated.

Thanks and regards

Keith

I'm currently evaluating the power of DynamicMethod to create an environment that based on user input can create more or less volatile methods.

The user input finally forming these methods may be buggy and throw exceptions.

There's a ton of questions that arise from this fact since it virtually seems to be impossible to support users where their error is. I would like to know how Dynamic method compilation internally works and what exactly the reasons are for omitting tables containing IL to native sequence point mappings in the .NET framework when working with dynamic methods.

I've also been experimenting with dynamic types concerning the scenario and found it unlikely that we can use them. Reason 1 is that the customer induced hosting platform for the project may not even support .NET 4.0, so getting rid again of the method code may be virtually impossible. Running different Appdomains is out of the question because of performance reasons. Reason 2: The memory overhead induced by loading tons of dynamic assemblies (and in worst case additionally AppDomains even) would kill the project anyway, since the dynamically generated code may be long living.

When evaluating the possibilities, I also noticed that is is impossible to step into the native machine code generated by the JIT-compiler for a dynamic method. I. e. you can't even have a look at the code you just generated from the IL you emitted. Why is that so?

Furthermore, emitting the same IL through ILGenerators from DynamicMethod and a MethodBuilder seems to lead to different byte code. The differences seem to be in the encoding of addresses for methods we called from within the emitted IL (In a test simply Console.WriteLine - two different encodings for the same OpCodes.Call). That was where I really lost it. I would like to have more information about what is happening there and why this should be the case at all.

The best way to realize the planned environment would be to precompile the dynamic method code just as the IL byte-array and serialize it to disk (or maybe even put it in a database) and only recompile when the user input file has changed. When the functionality is needed, the dynamicmethod would be quickly instantiated from the saved IL-byte-array. However, regarding the aforementioned issue, I am quite uncertain whether that would work. If the method addresses change on a restart of the environment, things would simply break. What do the emitted addresses depend on? The loaded assemblies? Their order of loading?

Next question: Is there a guarantee that the JITted code for a static 'DynamicType' method (generated by a MethodBuilder (maybe even with sequence points)) and a (also static) DynamicMethod will be the same if they based on the same IL code? Is there a possibility to access the Native-to-IL mapping of the MethodBuilder method? If so, this could be a workaround to map the native offset when an exception occurs (the only information you seem to be able to obtain for an exception in a DynamicMethod) to the IL that was generated.

I would appreciate direct feedback from Microsoft clarifying these points.

Following is some example code we used for evaluation:

using System;
using System.Text;
using System.Reflection;
using System.Reflection.Emit;
using System.Diagnostics;
using System.Diagnostics.SymbolStore;


namespace test
{
    class MainClass
    {
        internal delegate int callconv(int i);

        public static void Main(string[] args)
        {
            try
            {
                Assembly asm = AppDomain.CurrentDomain.GetAssemblies()[0];
                Module mod = asm.GetModules()[0];
                AssemblyBuilder asmb = AppDomain.CurrentDomain.DefineDynamicAssembly(new AssemblyName("test"), AssemblyBuilderAccess.RunAndSave);
                ModuleBuilder modb = asmb.DefineDynamicModule("test", @"TestAssembly.dll", true);
                ISymbolDocumentWriter doc = modb.DefineDocument(@"F:\fantasy\path\mymodule.xlg", Guid.Empty, Guid.Empty, Guid.Empty);
                TypeBuilder typb = modb.DefineType("test");
                MethodBuilder mb = typb.DefineMethod("tesmeth", MethodAttributes.Static | MethodAttributes.Public,
                            CallingConventions.Standard, typeof(int), new Type[] { typeof(int) });
                DynamicMethod x = new DynamicMethod("test", MethodAttributes.Public | MethodAttributes.Static, CallingConventions.Standard, typeof(int), new Type[] { typeof(int) }, modb, true);
                MethodInfo writelineint = typeof(Console).GetMethod("WriteLine", new Type[] { typeof(string), typeof(object) });
                ILGenerator ilg = x.GetILGenerator();
                //ilg.MarkSequencePoint(doc, 1, 2, 1, 32);
                ilg.Emit(OpCodes.Ldc_I4_4);
                ilg.Emit(OpCodes.Ldarg_0);
                ilg.Emit(OpCodes.Ldc_I4_2);
                ilg.Emit(OpCodes.Call, typeof(Console).GetMethod("WriteLine", new Type[] { typeof(int) }));
                ilg.Emit(OpCodes.Ldc_I4_3);
                ilg.Emit(OpCodes.Call, typeof(Console).GetMethod("WriteLine", new Type[] { typeof(int) }));
                //ilg.MarkSequencePoint(doc, 5, 2, 5, 32);
                ilg.Emit(OpCodes.Div);
                //ilg.MarkSequencePoint(doc, 7, 2, 7, 32);
                ilg.Emit(OpCodes.Box, typeof(int));
                ilg.Emit(OpCodes.Starg, 0);
                ilg.Emit(OpCodes.Ldstr, String.Intern("testres={0}"));
                ilg.Emit(OpCodes.Ldarg_0);
                ilg.Emit(OpCodes.Call, writelineint);
                ilg.Emit(OpCodes.Ldarg_0);
                ilg.Emit(OpCodes.Unbox, typeof(int));
                ilg.Emit(OpCodes.Ldobj, typeof(int));
                ilg.Emit(OpCodes.Ret);
                FieldInfo ilpropCode = ilg.GetType().GetField("m_ILStream", BindingFlags.Instance | BindingFlags.NonPublic);
                FieldInfo ilpropCodeLength = ilg.GetType().GetField("m_length", BindingFlags.Instance | BindingFlags.NonPublic);
                byte[] bytes = (byte[])ilpropCode.GetValue(ilg);
                int length = (int)ilpropCodeLength.GetValue(ilg);
                byte[] byIL = new byte[length];
                Array.Copy(bytes, byIL, length);
                object[] ats = x.GetCustomAttributes(true);

                callconv meth = (callconv)(x.CreateDelegate(typeof(callconv)));
                meth(2);
                Console.WriteLine("exec: {0}", meth(1));  // change to 0 here to get an exception for the dynamic method

                mb.DefineParameter(1, ParameterAttributes.In, "invalue");
                ilg = mb.GetILGenerator();
                ilg.MarkSequencePoint(doc, 2, 2, 2, 32);
                ilg.Emit(OpCodes.Ldc_I4_4);
                ilg.MarkSequencePoint(doc, 3, 2, 3, 32);
                ilg.Emit(OpCodes.Ldarg_0);
                ilg.Emit(OpCodes.Ldc_I4_2);
                ilg.Emit(OpCodes.Call, typeof(Console).GetMethod("WriteLine", new Type[] { typeof(int) }));
                ilg.Emit(OpCodes.Ldc_I4_3);
                ilg.Emit(OpCodes.Call, typeof(Console).GetMethod("WriteLine", new Type[] { typeof(int) }));
                ilg.MarkSequencePoint(doc, 5, 2, 5, 32);
                ilg.Emit(OpCodes.Div);
                ilg.MarkSequencePoint(doc, 7, 2, 7, 32);
                ilg.Emit(OpCodes.Box, typeof(int));
                ilg.Emit(OpCodes.Starg, 0);
                ilg.Emit(OpCodes.Ldstr, String.Intern("testres={0}"));
                ilg.Emit(OpCodes.Ldarg_0);
                ilg.Emit(OpCodes.Call, writelineint);
                ilg.Emit(OpCodes.Ldarg_0);
                ilg.Emit(OpCodes.Unbox, typeof(int));
                ilg.Emit(OpCodes.Ldobj, typeof(int));
                ilg.Emit(OpCodes.Ret);

                //mb.CreateMethodBody(byIL, byIL.Length);  // this will fail horribly
                Type typ = typb.CreateType();

                ilpropCode = ilg.GetType().GetField("m_ILStream", BindingFlags.Instance | BindingFlags.NonPublic);
                ilpropCodeLength = ilg.GetType().GetField("m_length", BindingFlags.Instance | BindingFlags.NonPublic);
                bytes = (byte[])ilpropCode.GetValue(ilg); // x.GetMethodBody().GetILAsByteArray(); // (byte[])getter.Invoke(d, null);
                length = (int)ilpropCodeLength.GetValue(ilg);
                byte[] byIL2 = new byte[length];
                Array.Copy(bytes, byIL2, length);

                if (!System.Linq.Enumerable.SequenceEqual<byte>(byIL, byIL2))
                {
                    Console.WriteLine("Code still differs!");
                }
                else
                {
                    Console.WriteLine("Equal code.");
                }

                MethodInfo mi = typ.GetMethod("tesmeth");

                callconv meth2 = (callconv)Delegate.CreateDelegate(typeof(callconv), mi);
                meth2(2);
                ats = meth2.Method.GetCustomAttributes(true);
                foreach (object o in ats)
                {
                    Console.WriteLine(o.ToString());
                }
                Console.WriteLine("exec2: {0}", meth2(0));


            }
            catch (Exception ex)
            {
                Console.WriteLine(ex.ToString());
                StackTrace trc = new StackTrace(ex, true);
                StackFrame sf = trc.GetFrames()[0];

                Console.WriteLine("meth:{2}, nat:{0}, il:{1}", sf.GetNativeOffset(), sf.GetILOffset(), sf.GetMethod().Name);
            }

            Console.ReadKey(true);
        }
    }
}

I have 

public class A{}
public class B
    {
        A a;
        public B(A a)
        {
            this.a =a;
        }
}

Now i am trying to invoke constructor of B using reflection.

A a = new A();
var type = typeof(B);
var ctr = type.GetConstructor(new []{typeof(A)});
var instance = ctr.Invoke(new Object[] { a });
Console.Out.Write(null != instance);

This runs fine in my local system however failing in the build server(using NAnt) and giving me an exception

System.Reflection.TargetInvocationException : Exception has been thrown by the target of an invocation.
  ----> System.InvalidProgramException : Common Language Runtime detected an invalid program.
at System.RuntimeMethodHandle._InvokeConstructor(Object[] args, SignatureStruct& signature, IntPtr declaringType)
at System.RuntimeMethodHandle.InvokeConstructor(Object[] args, SignatureStruct signature, RuntimeTypeHandle declaringType)
at System.Reflection.RuntimeConstructorInfo.Invoke(BindingFlags invokeAttr, Binder binder, Object[] parameters, CultureInfo culture)
at System.Reflection.ConstructorInfo.Invoke(Object[] parameters)
at Program..ctr(A a) in Program.cs:line 20
--InvalidProgramException

Any idea whats going wrong here and how can i debug further ?

Our web site is hosted in IIS with an AppPool set to run as a specific user account. When we directly refer to one of our managed assembly, no errors and the assembly is loaded ok. But the same assembly when referred to from the Mixed code (C++ code) throws the following exception

Could not load file or assembly 'MyDll, Version=0.0.0.0, Culture=neutral, PublicKeyToken=a348e42a78edea73' or one of its dependencies. The system cannot find the file specified.

Fusion Details show

=== Pre-bind state information ===
LOG: User = NT AUTHORITY\NETWORK SERVICE
LOG: DisplayName = CRC.XMLInterfaces, Version=0.0.0.0, Culture=neutral, PublicKeyToken=a348e42a78edea73
(Fully-specified)
LOG: Appbase = file:///C:/Windows/SysWOW64/inetsrv/
LOG: Initial PrivatePath = NULL
Calling assembly : (Unknown).
===
LOG: This bind starts in default load context.
LOG: No application configuration file found.
LOG: Using host configuration file: C:\Windows\Microsoft.NET\Framework\v4.0.30319\aspnet.config
LOG: Using machine configuration file from C:\Windows\Microsoft.NET\Framework\v4.0.30319\config\machine.config.
LOG: Post-policy reference: CRC.XMLInterfaces, Version=0.0.0.0, Culture=neutral, PublicKeyToken=a348e42a78edea73
LOG: Attempting download of new URL file:///C:/Windows/SysWOW64/inetsrv/CRC.XMLInterfaces.DLL.
LOG: Attempting download of new URL file:///C:/Windows/SysWOW64/inetsrv/CRC.XMLInterfaces/CRC.XMLInterfaces.DLL.
LOG: Attempting download of new URL file:///C:/Windows/SysWOW64/inetsrv/CRC.XMLInterfaces.EXE.
LOG: Attempting download of new URL file:///C:/Windows/SysWOW64/inetsrv/CRC.XMLInterfaces/CRC.XMLInterfaces.EXE.

Not sure why "AppBase" is set to inetsrv folder. Is there any way to force set the AppBase folder when loading from the mixed mode code. Same web site works perfectly fine on Windows Server 2008 R2. Problems started happening trying on Windows Server 2012.

Thanks

Hi,
I am trying to instrument methods in mscorlib.dll and put method calls into my own proxy methods into them. Now, I know I cannot add AssemblyRefs from mscorlib.dll to my proxy.dll because mscorlib does not like that in SingleDomain mode.
So, I want to merge all the content of proxy.dll (all typedefs, methods, methodbodies, ect) into mscorlib.dll. It is very similar to what has been discussed in http://social.msdn.microsoft.com/Forums/en-US/0338c6bf-81b5-44ff-8e43-8bdb9a44feb7/modify-systemexception-in-setilfunctionbody. 

I tried IMetaDataEmit::Merge() to inject all metadata into mscorlib.dll (at ModuleLoadStarted/Finished profiler callback). However, I could not get it to work yet. Documentation is very sparse on that topic. My problem is that I get a AccessViolation when I call MergeEnd:

  clr.dll!CMiniMdRW::GetFilterTable(void)     Unknown
  clr.dll!NEWMERGER::Merge(enum MergeFlags,enum CorRefToDefCheck)     Unknown
  clr.dll!RegMeta::MergeEnd(void)     Unknown

Here is the disassebly from the debugger:

CMiniMdRW::GetFilterTable:
    000007FEE8A7FA5C  mov         qword ptr [rsp+8],rbx 
    000007FEE8A7FA61  push        rdi 
    000007FEE8A7FA62  sub         rsp,20h 
    000007FEE8A7FA66  xor         ebx,ebx 
    000007FEE8A7FA68  mov         rdi,rcx 
--> 000007FEE8A7FA6B  cmp         qword ptr [rcx+18A0h],rbx 
    000007FEE8A7FA72  je          CMiniMdRW::GetFilterTable+0C5Eh (07FEE8A806BAh) 
    000007FEE8A7FA78  mov         rax,qword ptr [rdi+18A0h] 
    000007FEE8A7FA7F  mov         rbx,qword ptr [rsp+30h] 
    000007FEE8A7FA84  add         rsp,20h 
    000007FEE8A7FA88  pop         rdi 
    000007FEE8A7FA89  ret 
    000007FEE8A7FA8A  nop 
    000007FEE8A7FA8B  nop 

rcx is 0. That means that the this-pointer is NULL, if I interpret it correctly.

This is what I do in ModuleLoadFinished callback (simplified):

// open scope and IMetaDataImport for proxy.dll which is loaded as raw file
IMetaDataDispenserEx *pDisp;
CoCreateInstance(CLSID_CorMetaDataDispenser, NULL, CLSCTX_INPROC_SERVER, IID_IMetaDataDispenserEx, (void **)&pDisp);
IMetaDataImport   *pImport;
pDisp->OpenScopeOnMemory(pFileInMemory, fileSize, ofRead, IID_IMetaDataImport, (LPUNKNOWN*)&pImport);

// get IMetaDataEmit for mscorlib.dll
IMetaDataEmit2 * pEmit2;
corProfilerInfo->GetModuleMetaData(moduleId, ( DWORD)( ofWrite | ofRead), IID_IMetaDataEmit2, (IUnknown **) &pEmit2);

// do the merge
MyTokenMapper mapper; // dummy mapper which implements IMapToken
MyErrorHandler handler; // implements IUnknown, not sure what this one is supposed to do
emit->Merge(import, &mapper, &handler);
emit->MergeEnd(); // here, the crash happens

Can anyone tell me:
 1. Why does it crash? What can I do differently?
 2. I implemented IMapToken myself, but I do not get any callbacks to ::Map(). Isn't it supposed to give me a callback for each merge?
 3. Is it recommended to use ::Merge() at all? Or would it be more wise to use IMetaDataEmit::Define...() and do the merge one by one?

Thanks!
-Christoph