I'm writing some signal processing routine, using the PortAudio library. I'm using a
stucture which contains a pointer to float which is intended to be used as a buffer. I then pass it to an audio callback function.
My problem is that after callback processing is finished, my pointer has changed reference and thus cannot be freed. This is not such a big deal but the thing is that I don't understand when and how the pointer reference is changed and I'm getting a feel like I'm missing something important.
Here is a simplified version of the code :
typedef struct{
float* tmp;
//other stuff
} Data;
Data data;
data.tmp = NULL;
data.tmp = (float*) calloc(N,sizeof(float));// N is the size of the buffer
Pa_OpenDefaultStream(some args, //opens a PortAudio stream and passes tmp to callback
callback,
&data );
A stream is then started in another high priority thread and the callback is being executed as many times as needed. During callback tmp is being used as a ring buffer and is constantly being copied new data to.
static int callback(args,void* data){
Data* x = (Data*) tmp;
x->tmp = update();
}
where update() returns a pointer to a float which is initialized the same way as tmp is (calloc).
float* update(){
//do stuff
return m_tmp2;
}
float* m_tmp2 = (float*) calloc(N,sizeof(float));//same N as before
But after the stream is closed I get an error when calling free before quitting.
free(data.tmp);//throws a SIGABRT error
Some breakpoint debugging showed me that the reference of the pointer is being changed during the callback processing, but I don't get when and how it happens because everything else runs smoothly. It must be something during the callback execution, but I'm sure update() returns a pointer that is the same size as tmp. Or is it link with PortAudio ?
Please, any clues ?
Not really sure if I understand it right. You allocated the float (x.tmp) every time the callback function is called..
static int callback(args,void* data){
Data* x = (Data*) tmp;
x->tmp = update();
}
I assume the above is typo, you actually mean
static int callback(args,void* data){
Data* x = (Data*) data;
x->tmp = update();
}
Well, you're actually change the pointer value of tmp by assigning it update() because it's reallocate a new memory location in heap and changed the pointing location of the tmp..
float* update(){
//do stuff
return m_tmp2;
}
The data.tmp must have pointed to a new location every time the callback function is called.. So, I don't see why it doesn't behave as you described..
That's the correct behavior already.. Maybe I miss anything?
and maybe you should provide a mechanism to keep track of the buffer.. so all tmp (float *) you allocate for your circular buffer can be freed (not just the first one before the first callback is called..
Related
In C we should return pointer variable if define as 1- static stack 2- Heap-allocated 3-const address.
The question is:
Can argv** variable in main function be returned as a function return value without undefined behaviour?
I can't imagine a situation which requires to return argv** from a function, but you get it from main (int argc, char* argv[]) so yes. You can pass it as a parameter to a function and return it to another function. They're like const for you.
You can't return pointer to the data allocated temporarily for the function's own needs. int* F{ int A; return &A; } will not work. But you can set int* A=(int*)malloc(sizeof(int)); and return A;: it'll be your tiny wrapper around the malloc function. You can even receive int* A as a parameter and return A+1; the rule you mentioned in the beginning actually means "there are some restrictions of allocation methods if you're planning to use the pointer after the function exits". You can break the rule without returning the value (for example, you can store the poiner in a global variable and it'll become invalid when the function exits, like void F{ int A; GlobPtr=&A; } which will cause the same problems).
But your function does not allocate argv, it's allocated before your code starts. So you can even shift it by one value as you extract parameters one-by-one, the returned pointer *nextargv[] will also point to valid memory area.
Sorry for the post. I have researched this but..... still no joy in getting this to work. There are two parts to the question too. Please ignore the code TWI Reg code as its application specific I need help on nuts and bolts C problem.
So... to reduce memory usage for a project I have started to write my own TWI (wire.h lib) for ATMEL328p. Its not been put into a lib yet as '1' I have no idea how to do that yet... will get to that later and '2'its a work in progress which keeps getting added to.
The problem I'm having is with reading multiple bytes.
Problem 1
I have a function that I need to return an Array
byte *i2cBuff1[16];
void setup () {
i2cBuff1 = i2cReadBytes(mpuAdd, 0x6F, 16);
}
/////////////////////READ BYTES////////////////////
byte* i2cReadBytes(byte i2cAdd, byte i2cReg, byte i2cNumBytes) {
static byte result[i2cNumBytes];
for (byte i = 0; i < i2cNumBytes; i ++) {
result[i] += i2cAdd + i2cReg;
}
return result;
}
What I understand :o ) is I have declared a Static byte array in the function which I point to as the return argument of the function.
The function call requests the return of a pointer value for a byte array which is supplied.
Well .... it doesn't work .... I have checked multiple sites and I think this should work. The error message I get is:
MPU6050_I2C_rev1:232: error: incompatible types in assignment of 'byte* {aka unsigned char*}' to 'byte* [16] {aka unsigned char* [16]}'
i2cBuff1 = i2cReadBytes(mpuAdd, 0x6F, 16);
Problem 2
Ok say IF the code sample above worked. I am trying to reduce the amount of memory that I use in my sketch. By using any memory in the function even though the memory (need) is released after the function call, the function must need to reserve an amount of 'space' in some way, for when the function is called. Ideally I would like to avoid the use of static variables within the function that are duplicated within the main program.
Does anyone know the trade off with repeated function call.... i.e looping a function call with a bit shift operator, as apposed to calling a function once to complete a process and return ... an Array? Or was this this the whole point that C does not really support Array return in the first place.
Hope this made sense, just want to get the best from the little I got.
BR
Danny
This line:
byte *i2cBuff1[16];
declares i2cBuff1 as an array of 16 byte* pointers. But i2cReadBytes doesn't return an array of pointers, it returns an array of bytes. The declaration should be:
byte *i2cBuff1;
Another problem is that a static array can't have a dynamic size. A variable-length array has to be an automatic array, so that its size can change each time the function is called. You should use dynamic allocation with malloc() (I used calloc() instead because it automatically zeroes the memory).
byte* i2cReadBytes(byte i2cAdd, byte i2cReg, byte i2cNumBytes) {
byte *result = calloc(i2cNumBytes, sizeof(byte));
for (byte i = 0; i < i2cNumBytes; i ++) {
result[i] += i2cAdd + i2cReg;
}
return result;
}
Say I want to invoke a CUDA kernel, like this:
struct foo { int a; int b; float c; double d; }
foo arg;
// fill in elements of `arg` here
my_kernel<<<grid_size, block_size, 0, stream>>>(arg);
Assume that stream was previously created using a call to cudaStreamCreate(), so the above will execute asynchronously. I'm concerned about the required lifetime of arg.
Are the arguments to the kernel copied synchronously when I invoke it (so it would be safe for arg to go out of scope immediately), or are they copied asynchronously (so I need to ensure that it stays alive until the kernel runs)?
Arguments are copied synchronously at launch. The API exposes a call stack onto which execution parameters and function arguments are pushed in order, then a call finalises those arguments into a CUDA kernel launch on the drivers internal streams/command queues.
This process isn't documented, but as of CUDA 7.5, a runtime API kernel launch like this:
dot_product<<<1,n>>>(n, d_a, d_b);
becomes this:
(cudaConfigureCall(1, n)) ? (void)0 : (dot_product)(n, d_a, d_b);
where the host stub function dot_product is expanded into this:
void __device_stub__Z11dot_productiPfS_(int __par0, float *__par1, float *__par2)
{
if (cudaSetupArgument((void *)(char *)&__par0, sizeof(__par0), (size_t)0UL) != cudaSuccess) return;
if (cudaSetupArgument((void *)(char *)&__par1, sizeof(__par1), (size_t)8UL) != cudaSuccess) return;
if (cudaSetupArgument((void *)(char *)&__par2, sizeof(__par2), (size_t)16UL) != cudaSuccess) return;
{
volatile static char *__f __attribute__((unused)); __f = ((char *)((void ( *)(int, float *, float *))dot_product));
(void)cudaLaunch(((char *)((void ( *)(int, float *, float *))dot_product)));
};
}
void dot_product( int __cuda_0,float *__cuda_1,float *__cuda_2)
{
__device_stub__Z11dot_productiPfS_( __cuda_0,__cuda_1,__cuda_2);
}
cudaSetupArgument is the API call which is pushing arguments onto the call stack. Interestingly, this is actually deprecated in the API documentation for CUDA 7.5, even though the compiler is using it. I would, therefore, expect this to change in the future, but the idea will be the same.
The parameters of the kernel call are copied prior to execution, so the scope schould be of no concern. But please note that the size of all kernel parameters cannot exceed a maximum size in bytes. If you want larger structs or blobs of data you need to allocate the used memory on the device using cudaMalloc, then copy the content of the host struct to the device struct using cudaMemcpy and call the kernel with a pointer to the new device struct.
Your code would look something like this:
struct foo { int a; int b; float c; double d; }
foo arg;
foo *arg_d;
// fill in elements of `arg` here
cudaMalloc(&arg_d, sizeof(foo));
// check the allocation here
cudaMemcpy(arg_d, &arg, sizeof(foo), cudaMemcpyHostToDevice);
my_kernel<<<grid_size, block_size, 0, stream>>>(arg_d);
So I've been stuck on a memory problem for days now.
I have a multi-threaded program running with c++. I initialize a double* pointer.
From what I've read and previous programming experience, a pointer gets initialized to garbage. It will be Null if you initialize it to 0 or if you allocate memory that's too much for the program. For me, my pointer initialization, without allocation, gives me a null pointer.
A parser function I wrote is suppose to return a pointer to the array of parsed information. When I call the function,
double* data;
data = Parser.ReadCoordinates(&storageFilename[0]);
Now the returned pointer to the array should be set to data. Then I try to print something out from the array. I get memory corruption errors. I've ran gdb and it gives me a memory corruption error:
*** glibc detected *** /home/user/kinect/openni/Platform/Linux/Bin/x64-Debug/Sample-NiHandTracker: free(): corrupted unsorted chunks: 0x0000000001387f90 ***
*** glibc detected *** /home/user/kinect/openni/Platform/Linux/Bin/x64-Debug/Sample-NiHandTracker: malloc(): memory corruption: 0x0000000001392670 ***
Can someone explain to me what is going on? I've tried initializing the pointer as a global but that doesn't work either. I've tried to allocate memory but I still get a memory corruption error. The parser works. I've tested it out with a simple program. So I don't understand why it won't work in my other program. What am I doing wrong? I can also provide more info if needed.
Parser code
double* csvParser::ReadCoordinates(char* filename){
int x; //counter
int size=0; //
char* data;
int i = 0; //counter
FILE *fp=fopen(filename, "r");
if (fp == NULL){
perror ("Error opening file");
}
while (( x = fgetc(fp)) != EOF ) { //Returns the character currently pointed by the internal file position indicator
size++; //Number of characters in the csv file
}
rewind(fp); //Sets the position indicator to the beginning of the file
printf("size is %d.\n", size); //print
data = new char[23]; //Each line is 23 bytes (characters) long
size = (size/23) * 2; //number of x, y coordinates
coord = new double[size]; //allocate memory for an array of coordinates, need to be freed somewhere
num_coord = size; //num_coord is public
//fgets (data, size, fp);
//printf("data is %c.\n", *data);
for(x=0; x<size; x++){
fgets (data, size, fp);
coord[i] = atof(&data[0]); //convert string to double
coord[i+1] = atof(&data[11]); //convert string to double
i = i+2;
}
delete[] data;
fclose (fp);
return coord;
}
Corrupt memory occurs when you write outside the bound of an array or vector.
It's called heap underrun and overrun (depends on which side it's on).
The heap's allocation data gets corrupted, so the symptom you see is an exception in free() or new() calls.
You usually don't get an access violation because the memory is allocated and it belongs to you, but it's used by the heap's logic.
Find the place where you might be writing outside the bounds of an array.
This is a similar question to this SO post, which I have been unable to use to solve my problem. I have included some code here, which will hopefully help someone to bring home the message that the other posting was getting at.
I want to write a CLI/C++ method that can take a void pointer as a parameter and return the managed object (whose type I know) that it points to. I have a managed struct:
public ref struct ManagedStruct { double a; double b;};
The method I am trying to write, which takes a void pointer to the managed struct as a parameter and returns the struct.
ManagedStruct^ VoidPointerToObject(void* data)
{
Object^ result = Marshal::PtrToStructure(IntPtr(data), Object::typeid);
return (ManagedStruct^)result;
}
The method is called here:
int main(array<System::String ^> ^args)
{
// The instance of the managed type is created:
ManagedStruct^ myData = gcnew ManagedStruct();
myData->a = 1; myData->b = 2;
// Suppose there was a void pointer that pointed to this managed struct
void* voidPtr = &myData;
//A method to return the original struct from the void pointer
Object^ result = VoidPointerToObject(voidPtr);
return 0;
}
It crashes in the VoidPointerToObject method on calling PtrToStructure , with the error: The specified structure must be blittable or have layout information
I know this is an odd thing to do, but it is a situation I have encountered a few times, especially when unmanaged code makes a callback to managed code and passes a void* as a parameter.
(original explanation below)
If you need to pass a managed handle as a void* through native code, you should use
void* voidPtr = GCHandle::ToIntPtr(GCHandle::Alloc(o)).ToPointer();
// ...
GCHandle h = GCHandle::FromIntPtr(IntPtr(voidPtr));
Object^ result = h.Target;
h.Free();
(or use the C++/CLI helper class gcroot)
Marshal::PtrToStructure works on value types.
In C++/CLI, that means value class or value struct. You are using ref struct, which is a reference type despite use of the keyword struct.
A related problem:
void* voidPtr = &myData;
doesn't point to the object, it points to the handle.
In order to create a native pointer to data on the managed heap, you need to use pinning. For this reason, conversion between void* and Object^ isn't as useful as first glance suggests.