llvm: remove some workarounds in loadTruncate()

No longer needed with LLVM 21.
This commit is contained in:
Alex Rønne Petersen 2025-10-24 22:43:29 +02:00
parent feb05a716d
commit bebfdc3661

View File

@ -11399,16 +11399,6 @@ pub const FuncGen = struct {
const payload_llvm_ty = try o.lowerType(pt, payload_ty);
const abi_size = payload_ty.abiSize(zcu);
// llvm bug workarounds:
const workaround_explicit_mask = o.target.cpu.arch == .powerpc and abi_size >= 4;
const workaround_disable_truncate = o.target.cpu.arch == .wasm32 and abi_size >= 4;
if (workaround_disable_truncate) {
// see https://github.com/llvm/llvm-project/issues/64222
// disable the truncation codepath for larger than 32bits value - with this heuristic, the backend passes the test suite.
return try fg.wip.load(access_kind, payload_llvm_ty, payload_ptr, payload_alignment, "");
}
const load_llvm_ty = if (payload_ty.isAbiInt(zcu))
try o.builder.intType(@intCast(abi_size * 8))
else
@ -11422,14 +11412,7 @@ pub const FuncGen = struct {
else
loaded;
const anded = if (workaround_explicit_mask and payload_llvm_ty != load_llvm_ty) blk: {
// this is rendundant with llvm.trunc. But without it, llvm17 emits invalid code for powerpc.
const mask_val = try o.builder.intValue(payload_llvm_ty, -1);
const zext_mask_val = try fg.wip.cast(.zext, mask_val, load_llvm_ty, "");
break :blk try fg.wip.bin(.@"and", shifted, zext_mask_val, "");
} else shifted;
return fg.wip.conv(.unneeded, anded, payload_llvm_ty, "");
return fg.wip.conv(.unneeded, shifted, payload_llvm_ty, "");
}
/// Load a by-ref type by constructing a new alloca and performing a memcpy.