One issue to consider is that the char buffer might not be properly aligned for a struct (and, in this case, its int members num1 and num2). Depending on the platform and implementation, a 4-byte, 8-byte or 16-byte alignment might be required. For that reason, one alternative is to declare the buffer in terms of MyStruct initially and then access it via a char pointer:
MyStruct buffer[1024 / sizeof(MyStruct)];
char * cp = (char *) buffer;
// fill the buffer via cp
for (size_t i = 0; i < sizeof(buffer); ++i)
{
// do stuff with buffer[i]
}
If that approach is not possible, the buffer needs to be copied to another buffer with safe alignment; For example:
size_t n = sizeof(buffer) / sizeof(MyStruct);
MyStruct * p = (MyStruct *) malloc(n * sizeof(MyStruct));
if (!p) { exit(EXIT_FAILURE); }
memcpy(p, buffer, n * sizeof(MyStruct)); // copy buffer to p
for (size_t i = 0; i < n; ++i)
{
// do stuff with p[i]
}