================
@@ -3540,6 +3540,41 @@ def CIR_MemCpyOp : CIR_MemOp<"libc.memcpy"> {
// TODO: MemMoveOp
+//===----------------------------------------------------------------------===//
+// MemSetOp
+//===----------------------------------------------------------------------===//
+
+def CIR_MemSetOp : CIR_Op<"libc.memset"> {
+ let summary = "Equivalent to libc's `memset`";
+ let description = [{
+ Given the CIR pointer, `dst`, `cir.libc.memset` will set the first `len`
+ bytes of the memory pointed by `dst` to the specified `val`.
+
+ Examples:
+
+ ```mlir
+ // Set 2 bytes from a record to 0:
+ %2 = cir.const #cir.int<2> : !u32i
+ %3 = cir.const #cir.int<0> : !u32i
+ %zero = cir.cast integral %3 : !s32i -> !u8i
+ cir.libc.memset %2 bytes from %record set to %zero : !cir.ptr<!void>,
----------------
andykaylor wrote:
I wasn't sure about the order of changes (whether we are moving towards or away
from this sort of format), but I verified that this matches what we have for
the memcpy operation, though now looking again I find myself questioning the
use of "from" here. The memcpy operation looks like:
`cir.libc.memcpy %2 bytes from %arr to %record`
which is quite clear. It seems that if we are attempting this in memset, it
should be:
`cir.libc.memset %2 bytes at %record set to %zero`
Or perhaps
`cir.libc.memset %2 bytes at %record to %zero`
https://github.com/llvm/llvm-project/pull/178806
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
