Waking up Windows when in AwayMode - c++

I am using the Windows AwayMode to turn off the monitor and audio instead of going into sleep mode. This is working fine. How do I "wake up" the system when an event I want occurs? I can detect the event, but I don't know how to get the monitor back on and the system to appear awake again.
I have tried GetCursorPos() and SetCursorPos() to try move the cursor but that did not work.
I also tried the CreateWaitableTimer() and SetWaitableTimer() but that didn't work either. I set the fResume option to TRUE.
I also tried to turn off AwayMode with the PowerSetRequest() handle and setting it NULL. That also did not work.
I also have tried the SetThreadExecutionState() call with no luck. There is an AwayMode defined here as well. I tried to set that and clear it, but the monitor doesn't come back on.

I found a way to do it using SendInput() with a mouse movement. I also had to use the SetThreadExecutionState() to let the system know that a user is present otherwise it will go back to AwayMode in 2 seconds. Here is the code I used.
// Get the current position to ensure we put it back at the end
POINT pt;
GetCursorPos(&pt);
// Make a mouse movement
// Go to upper left corner (0,0)
INPUT input;
input.type = INPUT_MOUSE;
input.mi.mouseData = 0;
input.mi.dx = 0;
input.mi.dy = 0;
input.mi.dwFlags = MOUSEEVENTF_ABSOLUTE | MOUSEEVENTF_MOVE;
SendInput(1, &input, sizeof(input));
Sleep(5); // Just in case this is needed
// Go to lower right corner (65535,65535)
input.mi.dx = 65535;
input.mi.dy = 65535;
input.mi.dwFlags = MOUSEEVENTF_ABSOLUTE | MOUSEEVENTF_MOVE;
SendInput(1, &input, sizeof(input));
Sleep(5); // Just in case this is needed
// Restore to original
SetCursorPos(pt.x, pt.y);
// Now let the system know a user is present
DWORD state = SetThreadExecutionState(ES_DISPLAY_REQUIRED | ES_SYSTEM_REQUIRED);

Related

SendInput Coordinates Confusion

I am trying to understand how SendInput() works with mouse input. As an example, I have a dual monitor display. My 2160p monitor is "above" the 1440p primary monitor. I am trying to send mouse input to the bottom right corner of the above 2160 monitor. Unfortunately, the cursor just ends up in the bottom right corner of my 1440p primary monitor instead.
I know that the top left corner of my primary monitor is (0,0), and the bottom right is (2560,1440). From what I understand, all coordinates are in relation to the primary monitor. So shouldn't I just be able to subtract the resolution of the above monitor from the primary one to move to the exact same location on the above?
I have tried both adding and subtracting, but I keep ending up in the bottom corner of the primary monitor.
INPUT input;
input.type = INPUT_MOUSE;
input.mi.dx = (2560 - 3840) * (65536.f / (float)GetSystemMetrics(SM_CXSCREEN));
input.mi.dy = (1440 - 2160) * (65536.f / (float)GetSystemMetrics(SM_CYSCREEN));
input.mi.dwFlags = MOUSEEVENTF_ABSOLUTE | MOUSEEVENTF_MOVE;
input.mi.mouseData = 0;
input.mi.dwExtraInfo = NULL;
input.mi.time = 0;
SendInput(1, &input, sizeof(INPUT));
UPDATE: Here is one of my failed attempts using the RECT structure from GetMonitorInfo. I am probably making some sort of obvious mistake, but I can't figure it out.
DXGI_OUTPUT_DESC output_desc;
output->GetDesc(&output_desc);
MONITORINFOEX info;
info.cbSize = sizeof(info);
GetMonitorInfo(output_desc.Monitor, (LPMONITORINFO)&info);
INPUT input;
input.type = INPUT_MOUSE;
input.mi.dx = (info.rcMonitor.right) * (65536.f / (float)GetSystemMetrics(SM_CXVIRTUALSCREEN));
input.mi.dy = (info.rcMonitor.top) * (65536.f / (float)GetSystemMetrics(SM_CYVIRTUALSCREEN));
input.mi.dwFlags = MOUSEEVENTF_ABSOLUTE | MOUSEEVENTF_VIRTUALDESK| MOUSEEVENTF_MOVE;
input.mi.mouseData = 0;
input.mi.dwExtraInfo = NULL;
input.mi.time = 0;
SendInput(1, &input, sizeof(INPUT));

Why PostMessage with WM_LBUTTONDBLCLK does not work?

I run this code:
LPPOINT pp = new POINT;
GetCursorPos(pp);
while(1){
PostMessage(GetDesktopWindow(), WM_LBUTTONDBLCLK, 0, MAKELPARAM(pp->x, pp->y));
Sleep(1000);
}
It does not click on the point indicated by the cursor, but opens and closes the Start menu
.Please tell me what's wrong
Sending WM_LBUTTONDBLCLK to an arbitrary window handle or to the desktop will not simulate a mouse click.
You can use SendInput, however, which can simulate a mouse click given screen coordinates (not a window handle, or window or client coordinates). This code will simulate a left-button click at the current cursor position:
INPUT in[2]; // 0 = left dn, 1 = left up
ZeroMemory(in, sizeof(INPUT) * 2);
in[0].type = INPUT_MOUSE;
in[0].mi.dwFlags = MOUSEEVENTF_LEFTDOWN;
in[1].type = INPUT_MOUSE;
in[1].mi.dwFlags = MOUSEEVENTF_LEFTUP;
SendInput(2, in, sizeof(INPUT));
Note that you can also use mouse_event, but SendInput is preferred according to the official documentation on MSDN.

Simple program considered malware

I'm working on an alarm clock application. I've been testing some things with the following code:
FMOD_SYSTEM* sound;
FMOD_SOUND* son;
FMOD_System_Create(&sound);
FMOD_System_Init(sound, 32, FMOD_INIT_NORMAL, NULL);
FMOD_System_CreateSound(sound, "alarm.mp3", FMOD_CREATESAMPLE, 0, &son);
SYSTEMTIME time;
GetSystemTime(&time);
while (!(time.wHour==16 && time.wDayOfWeek==4)) {
Sleep(10000);
GetSystemTime(&time);
}
FMOD_System_PlaySound(sound, son, NULL, 0, NULL);
for (int i = 0; i < 100; i++) {
INPUT ip = { 0 };
ip.type = INPUT_KEYBOARD;
ip.ki.wVk = VK_VOLUME_UP;
SendInput(1, &ip, sizeof(INPUT));
ip.ki.dwFlags = KEYEVENTF_KEYUP;
SendInput(1, &ip, sizeof(INPUT));
}
The idea would be to pu it in the Startup folder so that it launches every time the user logs in, then waits until the specified hour. Then, it would play the sound file using the FMOD library and increase master volume. I know that my way of increasing volume is unorthodox, but I'm just testing for the moment. It works well if executed.
The problem is that my antivirus considers it malware. I've tried it under Avast and Bitdefender with the same results. It's a rather simple code that doesn't even manage files, how can it be considered a threat?
Thanks
Edit: Forgot to say that I've tested it with and without the master volume managing loop. Therefore, there should be a problem anywhere else...

Sending Mouse Click to DirectX Game Running as Admin

I have been building a gaming-related program that needs to send simulated input to the game (which is the top window on the screen and runs in fullscreen mode). After some struggling, I finally got mouse movements (cursor drag) and keyboard input working, but for some reason, the game will not respond to simulated mouse clicks.
I have tried the following:
#if TRUE // SendInput works for keyboard simulation and mouse drag, but not clicks:
INPUT mouse = {0};
mouse.type = INPUT_MOUSE;
mouse.mi.dwFlags = MOUSEEVENTF_LEFTDOWN;
SendInput(1, &mouse, sizeof(INPUT));
Sleep(100);
ZeroMemory(&mouse, sizeof INPUT);
mouse.type = INPUT_MOUSE;
mouse.mi.dwFlags = MOUSEEVENTF_LEFTUP;
SendInput(1, &mouse, sizeof(INPUT));
#else // tried this, but it did not work:
mouse_event(MOUSEEVENTF_LEFTDOWN, 0, 0, 0, 0);
Sleep(75);
mouse_event(MOUSEEVENTF_LEFTUP, 0, 0, 0, 0);
#endif
The first dilemma I faced was sending the messages in general, but I realized this issue was caused by game process privileges > sending application privileges. After this, I encountered another problem with keypresses only registering on certain game windows/screens, but after some searching, I was able to use scancodes to overcome this. For example:
void SendSpacePress(bool bHardwareLevel){
INPUT space = {0};
space.type = INPUT_KEYBOARD;
space.ki.time = 0;
space.ki.dwExtraInfo = 0;
if(!bHardwareLevel){
space.ki.wVk = VK_SPACE;
}else{
space.ki.wScan = 0x39; // physical keyboard scan code
}
space.ki.dwFlags = bHardwareLevel ? KEYEVENTF_SCANCODE : 0;
SendInput(1, &space, sizeof(INPUT));
Sleep(rand()%25 + 25);
space.ki.dwFlags = bHardwareLevel ? KEYEVENTF_SCANCODE | KEYEVENTF_KEYUP : KEYEVENTF_KEYUP;
SendInput(1, &space, sizeof(INPUT));
}
I have also been able to do mouse movement using INPUT like below:
void PanMouse(){
INPUT mouse = {0};
mouse.type = INPUT_MOUSE;
mouse.mi.time = 0;
mouse.mi.mouseData = 0;
mouse.mi.dwExtraInfo = 0;
mouse.mi.dwFlags = MOUSEEVENTF_MOVE;
mouse.mi.dx = rand()%10 -5;
mouse.mi.dy = rand()%10 -5;
SendInput(1, &mouse, sizeof(INPUT));
}
Now the big problem is the game refuses to register my simulated mouse clicks. I would preferably like a way to send the mouse inputs without having to dive into hooking. (The game has multiple client-side anti-cheat mechanisms, so I would venture to guess that any kind of foreign process hooking or DLL injection would trigger the protection.) From what I have read, I might need to write a custom driver for a simulated hardware mouse so the input comes at the kernel level. Not a preferable option, but if need be, sobeit. (And if need be, can anyone provide helpful information for this? I have never messed around with writing drivers, but there's a first time for everything I suppose.)
TL;DR: What do I need to do to get simulated mouse clicks to register in a game that seems to ignore non-hardware input? (I.e., how can I trick the game into thinking mouse clicks are legitimate similarly to using KEYEVENTF_SCANCODE when simulating keyboard input?)
Instead of sleeping between inputs, you need to inject a pair of mouse inputs together with the time member appropriately set to indicate "how long".
Here's something close to what you want. This will click the current cursor position with the left mouse simulated for 100ms.
INPUT inputs[2] = {};
inputs[0].type = INPUT_MOUSE;
inputs[0].mi.time = 0;
inputs[0].mi.dx = 0;
inputs[0].mi.dy = 0;
inputs[0].mi.dwFlags = MOUSEEVENTF_LEFTDOWN;
inputs[1].type = INPUT_MOUSE;
inputs[1].mi.time = 100;
inputs[1].mi.dx = 0;
inputs[1].mi.dy = 0;
inputs[1].mi.dwFlags = MOUSEEVENTF_LEFTUP;
SendInput(2, inputs, sizeof(INPUT));
The mouse event won't happen immediately. It gets "queued" to happen shortly afterwards.
You should use the same technique for your keyboard events. Don't call "Sleep".

Screen goes black when i use SendInput to send mouse cursor positions

I am using SendInput() to send relative positions of the mouse. First ill tel you what am doing.
i use my finger to move the mouse. So first i track my finger in a 640x480 image and get the absolute position in pixels with in the image.
then i send that position to the following method to generate relative mouse position commands using send input.
When the finger goes to the left boundary (xlim1) or the right boundary (xlim2) cursor keep scrolling horizontally to either left or right depending on which limit.
The issue is when i run the code and just when the cursor starts to move, screen goes to black.
when i comment the part else if(cx >= prevX && cx > xlim2){ .... } section, then it works.. (So the when finger point goes to right limit of the image it cursor keeps scrolling horizontally to the right. commented part enables the left horizontal scrolling).
bool first variable will be true if this is the first time, we capture the finger. Otherwise it is false.
void movMouse(int cx, int cy, bool first){
static int prevX = 0;
static int prevY = 0;
static int leftPrevX;
static int rightPrevX;
int mx,my;
if(first == true){
prevX = cx;
prevY = cy;
}
else{
mx = (cx - prevX);
my = (cy - prevY);
if(cx <= prevX && cx < xlim1){
mx = -20;
INPUT input;
input.type = INPUT_MOUSE;
input.mi.mouseData = 0;
input.mi.dx = -(mx);
input.mi.dy = (my);
input.mi.dwFlags = MOUSEEVENTF_MOVE;
SendInput(1, &input, sizeof(input));
}
else if(cx >= prevX && cx > xlim2){
mx = 20;
INPUT input;
input.type = INPUT_MOUSE;
input.mi.mouseData = 0;
input.mi.dx = -(mx);
input.mi.dy = (my);
input.mi.dwFlags = MOUSEEVENTF_MOVE;
SendInput(1, &input, sizeof(input));
}
else {
INPUT input;
input.type = INPUT_MOUSE;
input.mi.mouseData = 0;
input.mi.dx = -(mx);
input.mi.dy = (my);
input.mi.dwFlags = MOUSEEVENTF_MOVE;
SendInput(1, &input, sizeof(input));
}
prevX = cx;
prevY = cy;
}
}
Try
ZeroMemory(&input,sizeof(input));
also intialize all the variables including input.time
it worked for me :)
I ran into this same problem, even though I was calling ZeroMemory and doing everything else correctly. I was using input.mi.time to inform Windows of the spacing between the clicks, e.g. so double-click would work correctly. However I was getting the 'time' values from a remote computer. Because they differed from the local computer's time it caused Windows to invoke the screen saver! To workaround the problem I added some logic to detect the skew between the computers and bring the values somewhat in line with each other.
In summary: Make sure that input.mi.time is either zero or a value somewhat close to GetTickCount(). Using ZeroMemory to initialize the variable is an excellent suggestion.