| Issue |
150604
|
| Summary |
[DA] Insufficient validation of delinearization results
|
| Labels |
llvm:analysis
|
| Assignees |
|
| Reporter |
kasuga-fj
|
Input:
```llvm
define void @f(ptr %a, i64 %k) {
entry:
%cmp = icmp eq i64 %k, 0
br i1 %cmp, label %exit, label %loop
loop:
%i = phi i64 [ 0, %entry ], [ %i.next, %loop ]
%i.next = add i64 %i, 1
%subscript.0 = mul i64 %i, %k
%subscript.1 = mul i64 %i.next, %k
%idx.0 = getelementptr i8, ptr %a, i64 %subscript.0
%idx.1 = getelementptr i8, ptr %a, i64 %subscript.1
store i8 42, ptr %idx.0
store i8 42, ptr %idx.1
%cond.exit = icmp eq i64 %i.next, 10000
br i1 %cond.exit, label %exit, label %loop
exit:
ret void
}
```
Result:
```
$ opt -passes='print<da>' -disable-output test.ll
Printing analysis 'Dependence Analysis' for function 'f':
Src: store i8 42, ptr %idx.0, align 1 --> Dst: store i8 42, ptr %idx.0, align 1
da analyze - none!
Src: store i8 42, ptr %idx.0, align 1 --> Dst: store i8 42, ptr %idx.1, align 1
da analyze - consistent output [-1]!
Src: store i8 42, ptr %idx.1, align 1 --> Dst: store i8 42, ptr %idx.1, align 1
da analyze - none!
```
godbolt: https://godbolt.org/z/8Pc4Ev565
The dependencies between the two stores are reported as `[-1]`, which is incorrect if `%k` is negative, as it should be `[1]`. I think the delinearization check is insufficient .Executing `opt --pass='print<delinearization>'` yields the following output:
```
Delinearization on function f:
Inst: %idx.0 = getelementptr i8, ptr %a, i64 %subscript.0
In Loop with Header: loop
AccessFunction: 0
failed to delinearize
Inst: %idx.1 = getelementptr i8, ptr %a, i64 %subscript.1
In Loop with Header: loop
AccessFunction: 0
failed to delinearize
Inst: store i8 42, ptr %idx.0, align 1
In Loop with Header: loop
AccessFunction: {0,+,%k}<%loop>
Base offset: %a
ArrayDecl[UnknownSize][%k] with elements of 1 bytes.
ArrayRef[{0,+,1}<nuw><nsw><%loop>][0]
Inst: store i8 42, ptr %idx.1, align 1
In Loop with Header: loop
AccessFunction: {%k,+,%k}<%loop>
Base offset: %a
ArrayDecl[UnknownSize][%k] with elements of 1 bytes.
ArrayRef[{1,+,1}<nuw><nsw><%loop>][0]
Compiler returned: 0
```
godbolt: https://godbolt.org/z/9W6YWcMsG
The line `ArrayDecl[UnknownSize][%k] ...` indicates the estimated array size. Maybe we should reject such a result in DA if we cannot prove `%k` is non-negative.
![]()
_______________________________________________
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
