__arithmetic_fence reject sizeless SVE vector type
- Dominant language
- LLVM
- Stars
- 40.5k
- Forks
- 18.7k
- PR merge metrics
- PR metrics pending
Description
https://gcc.godbolt.org/z/PEvq1qnfo
```c++
// similar builtin in gcc accept sizeless sve
svfloat64_t test(svfloat64_t x) {
return __builtin_assoc_barrier(x);
}
```
```c++
// clang++ support fixed size vector
using fixed_t =
svfloat64_t __attribute__((arm_sve_vector_bits(256)));
fixed_t test(fixed_t x) {
return __arithmetic_fence(x);
}
```
```c++
// clang++ reject sizeless vector type
static_assert(__has_builtin(__arithmetic_fence));
svfloat64_t test(svfloat64_t x) {
return __arithmetic_fence(x);
}
```
error message:
```text
:30:32: error: invalid operand of type 'svfloat64_t' (aka '__SVFloat64_t') where floating, complex or a vector of such types is required
30 | return __arithmetic_fence(x);
| ^
1 error generated.
Compiler returned: 1
```
Contributor guide
Research direction
No repository file or test entry point is named. Start by reproducing the Godbolt examples and trace the __arithmetic_fence operand validation, comparing its handling with __builtin_assoc_barrier and fixed-size SVE vectors. Done means sizeless SVE vector operands are rejected with the intended diagnostic while supported fixed-size vectors continue to work.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp
- Domain
- compilers
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Active
- Clarity
- Mostly clear
- Newbie friendliness
- 45/100