Why the output of my program is the value of secret? - integer-overflow

#define INT_MAX 2147483647
#define INT_MIN (-INT_MAX-1)
int main()
{
int secret = 0x12345678;
int array[1] = {0};
printf("%d",array[INT_MIN+1]);
}
The code run without stack protector.
the question is about integer overflow, but i dont understand why it work because the value of INT_MIN + 1 is -2147483647. and if i want to print the value of secret, I need to write: array[-1]

Looking at your code, The printf returns 0 as it is an array of int, not the secret.
I think that there is an unpredictable behavior due to index outbound. You are declaring an array of 1 element but referring to an outbound index of the array. Maybe this post can help.

Related

Failure to compare strings with eBPF

When I run the following code I get an error.
#include <uapi/linux/utsname.h>
#include <linux/pid_namespace.h>
struct uts_namespace {
struct kref kref;
struct new_utsname name;
};
static __always_inline char * get_task_uts_name(struct task_struct *task){
return task->nsproxy->uts_ns->name.nodename;
}
int cmpNamespace(void *ctx) {
struct task_struct *task;
task = (struct task_struct *)bpf_get_current_task();
if (strcmp(get_task_uts_name(task),"namespace")==0){
...
}
return 0;
}
Error:
bpf: Failed to load program: Invalid argument
unknown opcode 00
processed 0 insns (limit 1000000) max_states_per_insn 0 total_states 0 peak_states 0 mark_read 0
HINT: The 'unknown opcode' can happen if you reference a global or static variable, or data in read-only section. For example, 'char *p = "hello"' will result in p referencing a read-only section, and 'char p[] = "hello"' will have "hello" stored on the stack.
But this works just fine
int cmpNamespace(void *ctx) {
char * test = "aaaa";
if (strcmp(test,"namespace")==0){
...
}
return 0;
}
Can anyone tell me why this is happening and how I could go about correcting it ?
I am using python bcc to hook the function.
Thanks!
The issue is that you are using strcmp. BPF programs cannot use functions from the libc.
Your second example probably works because the compiler is able to optimize it and remove the call to strcmp. Since both arguments are known at compile-time, there's no need to use strcmp to know if they are equal.
As pointed out by #Qeole in comments, you can use __builtin_memcmp() instead, since you know the size of one of your strings and are only trying to know if they are equal.

D language unsigned hash of string

I am a complete beginner with the D language.
How to get, as an uint unsigned 32 bits integer in the D language, some hash of a string...
I need a quick and dirty hash code (I don't care much about the "randomness" or the "lack of collision", I care slightly more about performance).
import std.digest.crc;
uint string_hash(string s) {
return crc320f(s);
}
is not good...
(using gdc-5 on Linux/x86-64 with phobos-2)
While Adams answer does exactly what you're looking for, you can also use a union to do the casting.
This is a pretty useful trick so may as well put it here:
/**
* Returns a crc32Of hash of a string
* Uses a union to store the ubyte[]
* And then simply reads that memory as a uint
*/
uint string_hash(string s){
import std.digest.crc;
union hashUnion{
ubyte[4] hashArray;
uint hashNumber;
}
hashUnion x;
x.hashArray = crc32Of(s); // stores the result of crc32Of into the array.
return x.hashNumber; // reads the exact same memory as the hashArray
// but reads it as a uint.
}
A really quick thing could just be this:
uint string_hash(string s) {
import std.digest.crc;
auto r = crc32Of(s);
return *(cast(uint*) r.ptr);
}
Since crc32Of returns a ubyte[4] instead of the uint you want, a conversion is necessary, but since ubyte[4] and uint are the same thing to the machine, we can just do a reinterpret cast with the pointer trick seen there to convert types for free at runtime.

How to Assign a single value to whole NSMutableArray?

I have an mutable array of say 20 objects. And it has values like #"TRUE",#"FALSE",#"TRUE"...
Now I want to reset the all values of array to #"FALSE". Means array having all values as #"FALSE" (20 times).
I know how to add, insert at index... But I want to know that How can I set whole array value to #"FALSE" in a sing line ??? `without using loop and replace object at index... ?
For example : is it possible
thArray = [[NSMutableArray alloc] initWithCapacity:20];
theArray = #"FALSE" ;
Thanks...
Can you use a C array? If so, you can use {0, 1} as C-equivalents of {FALSE, TRUE}, initializing a C array with:
unsigned short int cArray[20] = {0};
or:
static unsigned short int cArray[20]; /* all values are zeroes, or FALSEs */
This might be more efficient, instead of using an array of static NSString * const elements like you're doing now. Testing whether two integers are equivalent is usually faster than testing lexicographical equivalence of two strings, where your program will have to compare each string character by character.
To reset the contents of the array, you can use the C function memset():
memset(cArray, 0, 20*sizeof(unsigned short int)); /* set all values of cArray to 0 */
If you need a dynamically-sized array, use a pointer with calloc() and reset it with memset() as previously described. Just remember to free() the pointer afterwards, so that you don't get a memory leak:
unsigned short int *cArray = NULL;
size_t cArrayLength = 20; /* can be passed in as a value from another method, etc. */
cArray = calloc(cArrayLength, sizeof(unsigned short int)); /* values are initialized to 0 */
if (cArray) {
/* use cArray... */
*(cArray + 8) = 1; /* e.g., set ninth element with the value of 1 */
/* we don't need cArray any longer, so we free it */
free(cArray);
}
else
/* error */
If you must use Objective-C with NSString * or NSNumber * instances in an NSArray or NSMutableArray, there is no ready-made method for initialization and you'll need to use a loop or copy a pre-existing array, as described in Justin's answer. He is also correct that a method for creating and populating the array is a good idea, if you want to go in this direction.
If those were mutable strings inside the array you could do this in just one line:
[theArray makeObjectsPerformSelector:#selector(setString:) withObject:#"FALSE"];
Here's one way:
NSArray * initialized = ...array with fifty #"FALSE" values...;
NSMutableArray * a = [initialized mutableCopy];
... mutate a ...
[a setArray:initialized];
If you are actually dealing with bool values, C scalars will be faster (see Alex's answer).
Another alternative would be a CFMutableBitVector.
Finally, this would also be a good case for creating a function.
1) Either this
2) Or this
Otherwise without a loop there is no way.
only way i know is
for(int i = 0;i<[theArray count];i++)
{[theArray replaceObjectAtIndex:i withObject:#"FALSE"];}

Objective-C : Fowler–Noll–Vo (FNV) Hash implementation

I have a HTTP connector in my iPhone project and queries must have a parameter set from the username using the Fowler–Noll–Vo (FNV) Hash.
I have a Java implementation working at this time, this is the code :
long fnv_prime = 0x811C9DC5;
long hash = 0;
for(int i = 0; i < str.length(); i++)
{
hash *= fnv_prime;
hash ^= str.charAt(i);
}
Now on the iPhone side, I did this :
int64_t fnv_prime = 0x811C9DC5;
int64_T hash = 0;
for (int i=0; i < [myString length]; i++)
{
hash *= fnv_prime;
hash ^= [myString characterAtIndex:i];
}
This script doesn't give me the same result has the Java one.
In first loop, I get this :
hash = 0
hash = 100 (first letter is "d")
hash = 1865261300 (for hash = 100 and fnv_prime = -2128831035 like in Java)
Do someone see something I'm missing ?
Thanks in advance for the help !
In Java, this line:
long fnv_prime = 0x811C9DC5;
will yield in fnv_prime the numerical value -2128831035, because the constant is interpreted as an int, which is a 32-bit signed value in Java. That value is then sign-extended when written in a long.
Conversely, in the Objective-C code:
int64_t fnv_prime = 0x811C9DC5;
the 0x811C9DC5 is interpreted as an unsigned int constant (because it does not fit in a signed 32-bit int), with numerical value 2166136261. That value is then written into fnv_prime, and there is no sign to extend since, as far as the C compiler is concerned, the value is positive.
Thus you end up with distinct values for fnv_prime, which explains your distinct results.
This can be corrected in Java by adding a "L" suffix, like this:
long fnv_prime = 0x811C9DC5L;
which forces the Java compiler to interpret the constant as a long, with the same numerical value than what you get with the Objective-C code.
Incidentally, 0x811C9DC5 is not a FNV prime (it is not even prime); it is the 32 bit FNV "offset basis". You will get incorrect hash values if you use this value (and more hash collisions). The correct value for the 32 bit FNV prime is 0x1000193. See http://www.isthe.com/chongo/tech/comp/fnv/index.html
It is a difference in sign extension assigning the 32-bit value 0x811C9DC5 to a 64-bit var.
Are the characters in Java and Objective-c the same? NSString will give you unichars.

What does the & symbol mean in Objective-C?

What does the & symbol mean in Objective-C? I am currently looking at data constucts and am getting really confused by it.
I have looked around the web for it but have not found an answer at all. I know this is possibly a basic Objective-C concept, but I just can't get my head around it.
For example:
int *pIntData = (int *)&incomingPacket[0];
What is the code doing with incoming packet here?
& is the C address-of unary operator. It returns the memory address of its operand.
In your example, it will return the address of the first element of the incomingPacket array, which is then cast to an int* (pointer to int)
Same thing it means in C.
int *pIntData = (int *)&incomingPacket[0];
Basically this says that the address of the beginning of incomingPacket (&incomingPacket[0]) is a pointer to an int (int *). The local variable pIntData is defined as a pointer to an int, and is set to that value.
Thus:
*pIntData will equal to the first int at the beginning of incomingPacket.
pIntData[0] is the same thing.
pIntData[5] will be the 6th int into the incomingPacket.
Why do this? If you know the data you are being streamed is an array of ints, then this makes it easier to iterate through the ints.
This statement, If I am not mistaken, could also have been written as:
int *pIntData = (int *) incomingPacket;