https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101684
Bug ID: 101684
Summary: Different optimization levels lead to different output
results (-Wuninitialized)
Product: gcc
Version: 10.1.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: c++
Assignee: unassigned at gcc dot gnu.org
Reporter: drfeng08 at gmail dot com
Target Milestone: ---
The same code compiles as expected under O0 and gets an error under O3, the
compiler prompts a Warning (uninitialized in this function [-Wunitialized]), I
double-checked the relevant variables and it is impossible to have an
uninitialized case.
Minimal reproducible case
```
#include <iostream>
#include <cmath>
#include <cstddef>
#include <cstdint>
static const uint32_t ONE_BILLION = 1000000000;
static const uint32_t MURMUR3_32_SEED = 104729;
static uint32_t rotl32(uint32_t x, int8_t r) {
return (x << r) | (x >> (32 - r));
}
static uint32_t fmix32(uint32_t h) {
h ^= h >> 16;
h *= 0x85ebca6b;
h ^= h >> 13;
h *= 0xc2b2ae35;
h ^= h >> 16;
return h;
}
// modify from
https://github.com/aappleby/smhasher/blob/master/src/MurmurHash3.cpp
static uint32_t murmur_hash3_32(const void* key, int32_t len, uint32_t seed) {
const uint8_t* data = (const uint8_t*)key;
const int nblocks = len / 4;
uint32_t h1 = seed;
const uint32_t c1 = 0xcc9e2d51;
const uint32_t c2 = 0x1b873593;
const uint32_t* blocks = (const uint32_t*)(data + nblocks * 4);
for (int i = -nblocks; i; i++) {
uint32_t k1 = blocks[i];
k1 *= c1;
k1 = rotl32(k1, 15);
k1 *= c2;
h1 ^= k1;
h1 = rotl32(h1, 13);
h1 = h1 * 5 + 0xe6546b64;
}
const uint8_t* tail = (const uint8_t*)(data + nblocks * 4);
uint32_t k1 = 0;
switch (len & 3) {
case 3:
k1 ^= tail[2] << 16;
case 2:
k1 ^= tail[1] << 8;
case 1:
k1 ^= tail[0];
k1 *= c1;
k1 = rotl32(k1, 15);
k1 *= c2;
h1 ^= k1;
};
h1 ^= len;
h1 = fmix32(h1);
return h1;
}
class DecimalV2Value {
public:
DecimalV2Value() = default;
inline bool from_olap_decimal(int64_t int_value, int64_t frac_value) {
bool success = true;
bool is_negative = (int_value < 0 || frac_value < 0);
if (is_negative) {
int_value = std::abs(int_value);
frac_value = std::abs(frac_value);
}
if (frac_value > 9) {
frac_value = 9;
success = false;
}
_value = static_cast<__int128_t>(int_value) * ONE_BILLION + frac_value;
if (is_negative) _value = -_value;
return success;
}
private:
__int128_t _value;
};
struct DecimalFindOp {
uint32_t find_olap_engine(const void* data) {
DecimalV2Value value;
int64_t int_value = *(int64_t*)(data);
int32_t frac_value = *(int32_t*)((char*)data + sizeof(int64_t));
value.from_olap_decimal(int_value, frac_value);
return murmur_hash3_32((char*)&value, sizeof(DecimalV2Value), 0);
}
};
struct decimal12_t {
int64_t integer;
int32_t fraction;
} __attribute__((packed));
int main(int argc, char* argv[]) {
decimal12_t d12;
d12.integer = atoi(argv[0]);
d12.fraction= atoi(argv[1]);
DecimalFindOp op;
std::cout << op.find_olap_engine(&d12) << std::endl;
return 0;
}
```
> /opt/compiler/gcc-10/bin/g++ case.cpp -Wall -std=c++11 -O3 -g
```
case.cpp: In function 'int main(int, char**)':
case.cpp:35:18: warning: 'value' is used uninitialized in this function
[-Wuninitialized]
35 | uint32_t k1 = blocks[i];
| ^~
case.cpp:35:18: warning: '*((void*)& value +4)' is used uninitialized in this
function [-Wuninitialized]
case.cpp:35:18: warning: '*((void*)& value +8)' is used uninitialized in this
function [-Wuninitialized]
case.cpp:35:18: warning: '*((void*)& value +12)' is used uninitialized in this
function [-Wuninitialized]
```
> ./a.out 12 12
```
2167721464
```
if I use -O1
> /opt/compiler/gcc-10/bin/g++ case.cpp -Wall -std=c++11 -g
> ./a.out 12 12
output:
```
4203026776
```