| Issue |
64656
|
| Summary |
Regression lowering inline asm memory operand for RISCV64
|
| Labels |
new issue
|
| Assignees |
|
| Reporter |
jacobly0
|
```llvm
target triple = "riscv64"
@g = global i32 0
define void @f() {
call void asm "sw zero, $0", "=*m"(ptr elementtype(i32) @g)
ret void
}
```
```
$ llc-16 -version # 7cbf1a2591520c2491aa35339f227775f4d3adf6
LLVM (http://llvm.org/):
LLVM version 16.0.6
DEBUG build with assertions.
Default target: x86_64-unknown-linux-gnu
Host CPU: znver4
$ llc-16 repro.ll -o -
...
#APP
sw zero, 0(a0)
#NO_APP
...
$ llc-17 -version # 8f4dd44097c9ae25dd203d5ac87f3b48f854bba8
LLVM (http://llvm.org/):
LLVM version 17.0.0-rc2
DEBUG build with assertions.
Default target: x86_64-unknown-linux-gnu
Host CPU: znver4
$ llc-17 repro.ll -o -
...
#APP
error: invalid operand in inline asm: 'sw zero, $0'
<inline asm>:1:11: error: unknown operand
sw zero,
^
#NO_APP
...
```
![]()
_______________________________________________
llvm-bugs mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
