Current Situation:
I'm looking to try and measure the current CPU utilisation of my system in Hertz.
I've had a look at this answer which addresses my question, however I cannot seem to make the code work.
This is my current code in main.cpp (from the answer):
#include <Pdh.h>
#include <PdhMsg.h>
#include <Windows.h>
static PDH_HQUERY cpuQuery;
static PDH_HCOUNTER cpuTotal;
void init()
{
PDH_STATUS a = PdhOpenQuery(NULL, NULL, &cpuQuery);
PDH_STATUS i = PdhAddCounter(cpuQuery, L"\\Processor(_Total)\\% Processor Time", NULL, &cpuTotal);
PdhCollectQueryData(cpuQuery);
}
double getCurrentValue()
{
init();
PDH_FMT_COUNTERVALUE counterVal;
PdhCollectQueryData(cpuQuery);
PdhGetFormattedCounterValue(cpuTotal, PDH_FMT_DOUBLE, NULL, &counterVal);
return counterVal.doubleValue;
}
int main()
{
double CPUUsage = getCurrentValue();
}
The Problem:
The value returned from getCurrectValue() is zero.
Related Observations:
I have observed, that the values in a and i of type PDH_STATUS are both zero? I speculate that this could be related to my lack of a value in CPUUsage, although I'm not sure why the function wouldn't be returning properly into these values.
Additional information:
I haven't used PDH before.
It's a PdhAddCoutner() is language dependent. You should use PdhAddEnglishCounter() instead.
EDIT: And you should have a minimal delay (500ms) between the query ran in the init() and the query in getvalue().
Additional explanations:
Running your code on my windows 8.1 system, it turned out that in init(), the status i returned was PDH_CSTATUS_NO_OBJECT, which means that it didn't find the object "Processor(_Total)".
I first thought there was a typo, and verified the name of the object and the counter on Technet.
By curiosity, I ran the original windows command perfmon.exe and noticed that all the objects and counter were translated in my native language. Running your code using the native language name of the counter gave me the correct result.
About the timing constraint
Once the language issue addressed, debugging step by step I got meaningful values of my CPU usage. But once I removed the breakpoints, I got either 0 or 100.
So I dug a little further in this issue to find out on microsoft support that % rates require some minimal delays between two consecutive queries. Once I added a Sleep(1000) at exit of init, I got again meaningful CPU usage values.
/* Needed windows definitions by following header */
#include <windef.h>
/* Windows performance data helper */
#include <Pdh.h>
/* Storage of PDH query and associated cpu counter */
struct cpu_counter{
PDH_HQUERY query;
PDH_HCOUNTER counter;
};
/* Initialize query & counter */
int cpu_counter_init(struct cpu_counter* pcc)
{
if(PdhOpenQueryA(NULL, 0, &pcc->query) != ERROR_SUCCESS)
return -1;
if(PdhAddEnglishCounterA(pcc->query, "\\Processor(_Total)\\% Processor Time", 0, &pcc->counter) != ERROR_SUCCESS || PdhCollectQueryData(pcc->query) != ERROR_SUCCESS)
{
PdhCloseQuery(pcc->query);
return -2;
}
return 0;
}
/* Fetch data from query and evaluate current counter value */
int cpu_counter_get(struct cpu_counter* pcc)
{
PDH_FMT_COUNTERVALUE counter_val;
if(PdhCollectQueryData(pcc->query) != ERROR_SUCCESS || PdhGetFormattedCounterValue(pcc->counter, PDH_FMT_LONG, NULL, &counter_val) != ERROR_SUCCESS)
return -1;
return counter_val.longValue;
}
/* Close all counters of query and query itself at the end */
void cpu_counter_close(struct cpu_counter* pcc)
{
if(pcc->query != NULL)
{
PdhCloseQuery(pcc->query);
pcc->query = NULL;
}
}
No ugly statics, status checks, use of instances instead of global, minimal include, save even without unicode definition, solution of Christophe (plz rather upvote him than me) built-in.
Related
I'm trying to get a list of all "File Explorer" instances currently running. It's fairly straight forward getting a list that includes all instances, but I find myself running into a brick wall filtering that list to only "File Explorer" instances.
The following piece of code retrieves all Explorer instances, meaning both "File Explorer" and "Internet Explorer":
#include <comdef.h>
#include <ExDisp.h>
#include <ShlGuid.h>
#include <Windows.h>
#include <cstdio>
using _com_util::CheckError;
using std::puts;
_COM_SMARTPTR_TYPEDEF(IShellWindows, __uuidof(IShellWindows));
int main()
{
CheckError(::CoInitializeEx(nullptr, COINIT_APARTMENTTHREADED | COINIT_DISABLE_OLE1DDE));
// Acquire IShellWindows interface
IShellWindowsPtr spShellWindows{};
CheckError(spShellWindows.CreateInstance(CLSID_ShellWindows, nullptr, CLSCTX_LOCAL_SERVER));
// Request iterator
IUnknownPtr spEnum{};
CheckError(spShellWindows->_NewEnum(&spEnum));
IEnumVARIANTPtr spEnumVariant{};
CheckError(spEnum.QueryInterface(__uuidof(spEnumVariant), &spEnumVariant));
// Iterate over shell windows ...
while (true) {
variant_t var{};
// ... one element at a time
HRESULT hr = spEnumVariant->Next(1, &var, nullptr);
CheckError(hr);
// Iterator depleted?
if (hr == S_FALSE) break;
// Did we get the expected `IDispatch` interface?
if (var.vt != VT_DISPATCH) continue;
IDispatchPtr spDisp{};
spDisp.Attach(var.pdispVal, true);
puts("Got suspect; need ID");
// That was easy; now on to some real challenges
}
}
The obvious attempt
My first take at the problem was to just get rid of everything that isn't "File Explorer". Asking for the IWebBrowser2 interface would certainly only get an affirmative response from objects that actually are web browsers. Adding the following to the code above:
_COM_SMARTPTR_TYPEDEF(IWebBrowser2, __uuidof(IWebBrowser2));
// ...
int main()
{
// ...
IWebBrowser2Ptr spWebBrowser{};
hr = spDisp.QueryInterface(__uuidof(spWebBrowser), &spWebBrowser);
if (SUCCEEDED(hr)) puts("Implements IWebBrowser2");
// ...
After making the changes and running the code while an "Internet Explorer" instance is running produces the desired output. However, running the code while a "File Explorer" instance is running produces the same output! That's a surprise and a disappointment, all at the same time.
More robust, less useful
Exluding objects that can be identified as "not File Explorer" didn't work out. Let's try to only include objects that can be identified as "File Explorer" instead. That sounds even more obvious, but as we've learned, "obvious" and "not" go hand in hand when it comes to the Windows Shell.
I haven't actually implemented this, but the IShellWindows interface provides an Item method that can return only objects that match a particular ShellWindowTypeConstants (e.g. SWC_EXPLORER or SWC_BROWSER). Or return an object at a particular index in the window collection. But not BOTH!
So, yes, (potentially) more robust, but also less useful as it doesn't meet my requirements as soon as more than one instances of "File Explorer" are running. Bummer.
Circumstantial evidence
While neither of the above led anywhere, I started over and went on a full-blown investigation looking for hints. Since "File Explorer" browses the Shell namespace, there may be something to that account. The following outlines the approach, based on an article by Raymond Chen titled A big little program: Monitoring Internet Explorer and Explorer windows, part 1: Enumeration:
Starting from the IDispatch interface above, ask for a service with ID SID_STopLevelBrowser to get an IShellBrowser interface.
Call IShellBrowser::QueryActiveShellView to get an IShellView interface.
Ask the IShellView whether it implements something Shell-namespace-y, e.g. IPersistIDList.
If it does, conclude that we're holding a reference to a "File Explorer" instance.
This appears to produce the desired result, though it's not clear to me future-proof this is or when it stops working. Leaving aside how overly convoluted this appears, I'm concerned about its reliability.
Question
What's the recommended/robust/reliable way to identify all "File Explorer" instances in an IShellWindows collection? I will favor solutions based on official documentation, though I understand that this is the Windows Shell and there's next to no documentation at all.
Here's a possibility ...
#include <comdef.h>
#include <ExDisp.h>
#include <ShlGuid.h>
#include <Windows.h>
#include <cstdio>
#include <atlbase.h>
#include <string>
using _com_util::CheckError;
using std::puts;
using std::string;
_COM_SMARTPTR_TYPEDEF(IShellWindows, __uuidof(IShellWindows));
//nicked from StackOverflow
std::string ProcessIdToName(DWORD_PTR processId)
{
std::string ret;
HANDLE handle = OpenProcess(
PROCESS_QUERY_LIMITED_INFORMATION,
FALSE,
processId /* This is the PID, you can find one from windows task manager */
);
if (handle)
{
DWORD buffSize = 1024;
CHAR buffer[1024];
if (QueryFullProcessImageNameA(handle, 0, buffer, &buffSize))
{
ret = buffer;
}
else
{
printf("Error GetModuleBaseNameA : %lu", GetLastError());
}
CloseHandle(handle);
}
else
{
printf("Error OpenProcess : %lu", GetLastError());
}
return ret;
}
int main()
{
CheckError(::CoInitializeEx(nullptr, COINIT_APARTMENTTHREADED | COINIT_DISABLE_OLE1DDE));
// Acquire IShellWindows interface
IShellWindowsPtr spShellWindows{};
CheckError(spShellWindows.CreateInstance(CLSID_ShellWindows, nullptr, CLSCTX_LOCAL_SERVER));
// Request iterator
IUnknownPtr spEnum{};
CheckError(spShellWindows->_NewEnum(&spEnum));
IEnumVARIANTPtr spEnumVariant{};
CheckError(spEnum.QueryInterface(__uuidof(spEnumVariant), &spEnumVariant));
// Iterate over shell windows ...
while (true) {
variant_t var{};
// ... one element at a time
HRESULT hr = spEnumVariant->Next(1, &var, nullptr);
CheckError(hr);
// Iterator depleted?
if (hr == S_FALSE) break;
// Did we get the expected `IDispatch` interface?
if (var.vt != VT_DISPATCH) continue;
IDispatchPtr spDisp{};
spDisp.Attach(var.pdispVal, true);
puts("Got suspect; need ID");
// That was easy; now on to some real challenges
CComPtr<IWebBrowser2> lpWB;
spDisp->QueryInterface(&lpWB);
SHANDLE_PTR hWnd{ 0 };
lpWB->get_HWND(&hWnd);
if (hWnd)
{
DWORD pid = 0;
GetWindowThreadProcessId((HWND)hWnd, &pid);
if (pid != 0)
{
puts("pid");
auto s = ProcessIdToName((DWORD_PTR)pid);
puts(s.c_str());
}
}
}
}
Our program has a piece of code that calculates the list of computers on our local network. It uses the Windows Networking API (WNetOpenEnum/WNetEnumResource) to unwind the network. For many years, the resulting list was identical to the one that can be seen in Windows Explorer under the "Network" entry. However, recently we have noticed that the same code returns an empty list. During debugging I found that WNetOpenEnum returns error 1231 (ERROR_NETWORK_UNREACHABLE) when it is called for the "Microsoft Windows Network" under the root node.
I have to mention, though I'm pretty sure it has nothing to do with the matter, that the network unwinding is done multithreaded, to avoid possible delays in the main GUI thread. Each time a node of type RESOURCEUSAGE_CONTAINER is encountered, a new worker thread is launched. The thread function calls the following procedure:
DWORD WINAPI EnumNetwork(NETRESOURCE_M* lpNR)
{
const int BUF_SIZE = 16384; // 16K is a good size.
HANDLE hEnum;
DWORD Result;
// Call the WNetOpenEnum function to begin the enumeration.
Result = ::WNetOpenEnum(RESOURCE_GLOBALNET, // all network
RESOURCETYPE_ANY, // all resource types
0, // enumerate all
(LPNETRESOURCE)lpNR,// parent resource
&hEnum); // enumeration handle
if (Result != NO_ERROR) // -> for "Microsoft Windows Network" Result = 1231
return Result;
std::vector<std::wstring> SrvList;
// Allocate buffer for enumeration.
LPNETRESOURCE lpEnumNR = (LPNETRESOURCE)new char[BUF_SIZE];
if (lpEnumNR == 0)
Result = ERROR_OUTOFMEMORY;
else
{
while (1)
{
::ZeroMemory(lpEnumNR, BUF_SIZE); // Initialize the buffer.
DWORD NumEntries = -1; // Enumerate all entries.
DWORD BufSize = BUF_SIZE;
// Call WNetEnumResource to continue the enumeration.
Result = ::WNetEnumResource(hEnum, // enumeration handle
&NumEntries,// number of entries to enumerate
lpEnumNR, // array of resources to return
&BufSize); // buffer size
if (Result == NO_ERROR)
{
// If the call succeeds, loop through the array.
for (unsigned i = 0; i < NumEntries; ++i)
{
if (lpEnumNR[i].dwDisplayType == RESOURCEDISPLAYTYPE_SERVER)
{
// Collect servers.
LPCWSTR SrvName = lpEnumNR[i].lpRemoteName;
if (PathHelpers::IsFullPath(SrvName))
SrvList.push_back(SrvName);
}
else if ((lpEnumNR[i].dwUsage & RESOURCEUSAGE_CONTAINER) &&
lpEnumNR[i].lpRemoteName != 0)
{
TCHAR PathBuf[1024] = {0};
if (lpNR && lpNR->Path)
{
_tcscpy(PathBuf, lpNR->Path);
::PathAddBackslash(PathBuf);
}
_tcscat(PathBuf, lpEnumNR[i].lpRemoteName);
if (RegisterServer(PathBuf))
{
// Start new thread for recursive enumeration.
NETRESOURCE_M* lpChildNR = DeepCopyNR(&lpEnumNR[i], PathBuf);
ExploreNetwork(lpChildNR); // -> this starts a worker thread
}
else
{
GetLogger().LogMessage(
_T("Cycles found while unwinding network: %s"), PathBuf);
}
}
}
}
else
{
if (Result == ERROR_NO_MORE_ITEMS)
Result = NO_ERROR;
break;
}
} // end while
delete [] (char*)lpEnumNR;
} // end if
::WNetCloseEnum(hEnum);
if (!SrvList.empty())
NotifyServerAdded(SrvList);
return Result;
}
where NETRESOURCE_M is the structure
struct NETRESOURCE_M
{
NETRESOURCE NR;
LPTSTR Path;
};
Trying to figure out what could have caused such a sudden change in behavior, I found in Google that a few years ago Microsoft disabled the SMB1 protocol, which could affect Network Discovery. However, I can't believe they could have damaged their own API without saying a word in the documentation.
EDIT: At the same time, Windows Explorer has a bunch of computers under its "Network" node. In the network settings, the network type is "Domain", and the network discovery is ON. Services "Function Discovery Provider Host" and "Function Discovery Resources Publication" are running. Windows OS build is 19042.685.
Edit 2: The Sysinternals' "ShareEnum" tool also fails with the error: "No domains or workgroups where found on your network". Because of this, and also because some time ago our company moved all of its computers to a different network, I got the feeling that the problem is in the network configuration. Such as though the network is declared as "Domain", the computers were not enrolled to this domain. I do not understand much in that, but something like this.
I tried creating a mutant in my process that closes the process when it is created more than once with this code:
HANDLE m;
OBJECT_ATTRIBUTES attr;
attr.Length = sizeof(OBJECT_ATTRIBUTES);
UNICODE_STRING str;
RtlInitUnicodeString(&str, L"123");
attr.ObjectName = &str;
long stats;
if (!(stats = NtCreateMutant(&m, MUTANT_ALL_ACCESS, &attr, false))) {
if (0xC0000035 == stats) {
return false;
}
return true;
}
All of the functions have the correct address
When I remove the code attr.ObjectName = &str; the mutant handle is not null, but the second process does not exit since the NTSTATUS is 0
When I keep the code, NTSTATUS returns STATUS_OBJECT_PATH_SYNTAX_BAD, which I have no idea what has to do with NtCreateMutant
I want to know if it is even possible to do this using NtCreateMutex in usermode and to fix these issues or if I should just stop trying and just use CreateMutex (I dont want to use it for no reason other than I wanted to try using NtCreateMutex)
I know it's 1 year late but considering how many questions are on stack overflow, better later than never.
I`m quite surprised why #Raymond Chen made that comment, because you can use undocumented API's just fine. They are everywhere in Windows, and they hardly changed for the last decade.
They are also used in literally every Windows kernel driver possible, so they are not really "undocumented" and "unsupported".
First of all, that's not how you initialise OBJECT_ATTRIBUTES; if you want to do it that way you would do:
OBJECT_ATTRIBUTES ObjectAttributes; = { sizeof(ObjectAttributes) };
The recommended and official way is to use the InitializeObjectAttributes macro.
Second, the mutex name cannot be "123" inside the NT namespace. That's why you get STATUS_OBJECT_PATH_SYNTAX_BAD.
Third, your code is just ugly written and the usage is bad. What you want to create (from what it sounds) is a "process barrier".
I will give you a working full sample code:
BOOL CreateProcessBarrier()
{
NTSTATUS Result;
UNICODE_STRING MutantName;
OBJECT_ATTRIBUTES ObjectAttributes;
HANDLE MutantHandle;
WCHAR MutantNameNT[] = L"\\??\\40691290-71d5-45bc-b86a-e714496f4bf2";
WCHAR MutantNameWin[] = L"40691290-71d5-45bc-b86a-e714496f4bf2";
RtlInitUnicodeString(&MutantName, MutantNameNT);
InitializeObjectAttributes(&ObjectAttributes, &MutantName, 0, NULL, NULL);
// ALL OK
if (0 <= (Result = ZwCreateMutant(&MutantHandle, MUTANT_ALL_ACCESS, &ObjectAttributes, TRUE)))
{
// Banana
return TRUE;
}
// Already exists
else if (Result == STATUS_OBJECT_NAME_COLLISION || Result == STATUS_OBJECT_NAME_EXISTS)
{
// Que paso ?
return FALSE;
}
// Mutex creation failed, fallback to winapi
else
{
// Papaya
CreateMutexW(NULL, TRUE, MutantNameWin);
if (GetLastError() == ERROR_ALREADY_EXISTS)
{
return FALSE;
}
}
return TRUE;
}
The function above has support to fallback to normal WinApi in case it fails.
Usage is simple as:
Process A:
CreateProcessBarrier();
Process B,C,D,etc:
if(!CreateProcessBarrier()) ... do whatever.
Enjoy.
I have built a mp3 decoder directshow filter for win CE , and I want to measure the performance of the decoder. I found two macros from msdn site, https://msdn.microsoft.com/en-IN/library/ms932254.aspx which is declared in the measure.h header file in base classes.
It is explained in the measure.h file that these macros will expand to nothing unless macro PERF is defined. but once I enable the macro, I get link error
"LNK2019: unresolved external symbol Msr_Start() referneced in
function function "public: virtual long_cdecl
CMP3Decoder::Recieve(Struct IMediaSample
*)"(?Recieve#CMP3Decoder##UAAJPAUIMediaSample###Z)
I tried to dump the symbols in strmbase.lib, but I couldn't find any symbol name Msr_Start in it. also I searched the whole base classes folder source code.
where can I find the definition for these functions?
Or is there any other ways to measure the performance of the filter?
CMP3Decoder::recieve() function is as follows
HRESULT CMP3Decoder::Receive(IMediaSample *pSample)
{
HRESULT hr;
ASSERT(pSample);
if(pSample == NULL || m_MP3DecHandle == NULL)
{
return E_FAIL;
}
ASSERT (m_pOutput != NULL) ;
// Start timing the transform (if PERF is defined)
MSR_START(m_idTransform);
// have the derived class transform the data
hr = MP3StartDecode(pSample);//, pOutSample);
// Stop the clock and log it (if PERF is defined)
MSR_STOP(m_idTransform);
if (FAILED(hr)) {
//DbgLog((LOG_TRACE,1,TEXT("Error from transform")));
} else {
// the Transform() function can return S_FALSE to indicate that the
// sample should not be delivered; we only deliver the sample if it's
// really S_OK (same as NOERROR, of course.)
if (hr == NOERROR) {
//hr = m_pOutput->Deliver(pOutSample);
m_bSampleSkipped = FALSE; // last thing no longer dropped
} else {
// S_FALSE returned from Transform is a PRIVATE agreement
// We should return NOERROR from Receive() in this cause because returning S_FALSE
// from Receive() means that this is the end of the stream and no more data should
// be sent.
if (S_FALSE == hr) {
// Release the sample before calling notify to avoid
// deadlocks if the sample holds a lock on the system
// such as DirectDraw buffers do
//pOutSample->Release();
m_bSampleSkipped = TRUE;
if (!m_bQualityChanged) {
NotifyEvent(EC_QUALITY_CHANGE,0,0);
m_bQualityChanged = TRUE;
}
return NOERROR;
}
}
}
// release the output buffer. If the connected pin still needs it,
// it will have addrefed it itself.
//pOutSample->Release();
return hr;
}
According to MSDN, you need to link to Strmiids.lib.
Or is there any other ways to measure the performance of the filter?
To measure performance of a filter, I typically insert a custom trans-in-place filter before and after the filter to be measured. The trans-in-place filter outputs sample times and current time in high resolution to a log file. You can calculate filter processing time by subtracting the before current times from the after and averaging those, etc. Also, the file-io should only be done after stopping the graph as you don't want to interfere with the measurement itself.
Update:
Dumping the symbols in Strmiids.lib seems to confirm that Msr_xxx functions are not defined in Strmiids.lib. Looks like the MSDN article is incorrect.
How can you find out if an executable is running on Windows given the process name, e.g. program.exe?
The C++ standard library has no such support. You need an operating system API to do this. If this is Windows then you'd use CreateToolhelp32Snapshot(), followed by Process32First and Process32Next to iterate the running processes. Beware of the inevitable race condition, the process could have exited by the time you found it.
I just created one using Hans suggestion. Works like a champ!
Oh and here is the basic code.
Please you will have to add CStrings sAppPath and sAppName.
StartProcess is a small function that uses CreateProcess and returns the PID(not used here). You will need to replace it.
This is not a complete program, just the code to find if the program is running using Hans suggestion. A fun test is to set the path to c:\windows\ and the app to notepad.exe and set it for 10 seconds.
#include <tlhelp32.h>
PROCESSENTRY32 pe32 = {0};
HANDLE hSnap;
int iDone;
int iTime = 60;
bool bProcessFound;
while(true) // go forever
{
hSnap = CreateToolhelp32Snapshot(TH32CS_SNAPPROCESS,0);
pe32.dwSize = sizeof(PROCESSENTRY32);
Process32First(hSnap,&pe32); // Can throw away, never an actual app
bProcessFound = false; //init values
iDone = 1;
while(iDone) // go until out of Processes
{
iDone = Process32Next(hSnap,&pe32);
if (strcmp(pe32.szExeFile,sAppName) == 0) // Did we find our process?
{
bProcessFound = true;
iDone = 0;
}
}
if(!bProcessFound) // if we didn't find it running...
{
startProcess(sAppPath+sAppName,""); // start it
}
Sleep(iTime*1000); // delay x amount of seconds.
}
Assumptions: since you mention '.exe', you want this for some flavor of Windows. You want to write a program in C++ to determine whether a program with a particular executable name is running (regardless of the language used to implement the target program).
Enumerate the running processes using either the Toolhelp API or the process status API. Compare the name of the executable for each running process to the one you're looking for (and be aware that there may be more than one process with that executable name).
hProcessInfo = OpenProcess( PROCESS_ALL_ACCESS, FALSE, pe32.th32ProcessID );
do{
if(strcmp(pe32.szExeFile,"process.exe") == 0)
{
processfound = true;
break;
}
}while( Process32Next( hProcessSnap, &pe32 ) );
If you don't want to get process detail from code just press Ctrl+Alt+Del and check process list.