mirror/zig
1
0
Fork 0
mirror of https://github.com/ziglang/zig.git synced 2026-01-21 06:45:24 +00:00
Code Issues Packages Projects Releases Wiki Activity

llvm/cbe: fix signed @mod/@divFloor computations

Browse Source 
Closes #15636
This commit is contained in:
Jacob Young 2023-05-10 01:56:48 -04:00
parent 31a13ce2e3
commit 2e6a6d7564
3 changed files with 62 additions and 39 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

39
lib/zig.h
View File

@ -487,14 +487,14 @@ typedef ptrdiff_t intptr_t;
zig_basic_operator(uint##w##_t, div_floor_u##w, /) \
\
static inline int##w##_t zig_div_floor_i##w(int##w##_t lhs, int##w##_t rhs) { \
return lhs / rhs - (((lhs ^ rhs) & (lhs % rhs)) < INT##w##_C(0)); \
return lhs / rhs + (lhs % rhs != INT##w##_C(0) ? zig_shr_i##w(lhs ^ rhs, UINT8_C(w) - UINT8_C(1)) : INT##w##_C(0)); \
} \
\
zig_basic_operator(uint##w##_t, mod_u##w, %) \
\
static inline int##w##_t zig_mod_i##w(int##w##_t lhs, int##w##_t rhs) { \
int##w##_t rem = lhs % rhs; \
return rem + (((lhs ^ rhs) & rem) < INT##w##_C(0) ? rhs : INT##w##_C(0)); \
return rem + (rem != INT##w##_C(0) ? rhs & zig_shr_i##w(lhs ^ rhs, UINT8_C(w) - UINT8_C(1)) : INT##w##_C(0)); \
} \
\
static inline uint##w##_t zig_shlw_u##w(uint##w##_t lhs, uint8_t rhs, uint8_t bits) { \
@ -1078,7 +1078,7 @@ static inline int64_t zig_bit_reverse_i64(int64_t val, uint8_t bits) {
uint##w##_t temp = val - ((val >> 1) & (UINT##w##_MAX / 3)); \
temp = (temp & (UINT##w##_MAX / 5)) + ((temp >> 2) & (UINT##w##_MAX / 5)); \
temp = (temp + (temp >> 4)) & (UINT##w##_MAX / 17); \
return temp * (UINT##w##_MAX / 255) >> (w - 8); \
return temp * (UINT##w##_MAX / 255) >> (UINT8_C(w) - UINT8_C(8)); \
} \
\
zig_builtin_popcount_common(w)
@ -1298,15 +1298,6 @@ static inline zig_i128 zig_rem_i128(zig_i128 lhs, zig_i128 rhs) {
return lhs % rhs;
}
static inline zig_i128 zig_div_floor_i128(zig_i128 lhs, zig_i128 rhs) {
return zig_div_trunc_i128(lhs, rhs) - (((lhs ^ rhs) & zig_rem_i128(lhs, rhs)) < zig_make_i128(0, 0));
}
static inline zig_i128 zig_mod_i128(zig_i128 lhs, zig_i128 rhs) {
zig_i128 rem = zig_rem_i128(lhs, rhs);
return rem + (((lhs ^ rhs) & rem) < zig_make_i128(0, 0) ? rhs : zig_make_i128(0, 0));
}
#else /* zig_has_int128 */
static inline zig_u128 zig_not_u128(zig_u128 val, uint8_t bits) {
@ -1394,20 +1385,26 @@ static zig_i128 zig_rem_i128(zig_i128 lhs, zig_i128 rhs) {
return __modti3(lhs, rhs);
}
static inline zig_i128 zig_mod_i128(zig_i128 lhs, zig_i128 rhs) {
zig_i128 rem = zig_rem_i128(lhs, rhs);
return zig_add_i128(rem, ((lhs.hi ^ rhs.hi) & rem.hi) < INT64_C(0) ? rhs : zig_make_i128(0, 0));
}
static inline zig_i128 zig_div_floor_i128(zig_i128 lhs, zig_i128 rhs) {
return zig_sub_i128(zig_div_trunc_i128(lhs, rhs), zig_make_i128(0, zig_cmp_i128(zig_and_i128(zig_xor_i128(lhs, rhs), zig_rem_i128(lhs, rhs)), zig_make_i128(0, 0)) < INT32_C(0)));
}
#endif /* zig_has_int128 */
#define zig_div_floor_u128 zig_div_trunc_u128
static inline zig_i128 zig_div_floor_i128(zig_i128 lhs, zig_i128 rhs) {
zig_i128 rem = zig_rem_i128(lhs, rhs);
int64_t mask = zig_or_u64((uint64_t)zig_hi_i128(rem), zig_lo_i128(rem)) != UINT64_C(0)
? zig_shr_i64(zig_xor_i64(zig_hi_i128(lhs), zig_hi_i128(rhs)), UINT8_C(63)) : INT64_C(0);
return zig_add_i128(zig_div_trunc_i128(lhs, rhs), zig_make_i128(mask, (uint64_t)mask));
}
#define zig_mod_u128 zig_rem_u128
static inline zig_i128 zig_mod_i128(zig_i128 lhs, zig_i128 rhs) {
zig_i128 rem = zig_rem_i128(lhs, rhs);
int64_t mask = zig_or_u64((uint64_t)zig_hi_i128(rem), zig_lo_i128(rem)) != UINT64_C(0)
? zig_shr_i64(zig_xor_i64(zig_hi_i128(lhs), zig_hi_i128(rhs)), UINT8_C(63)) : INT64_C(0);
return zig_add_i128(rem, zig_and_i128(rhs, zig_make_i128(mask, (uint64_t)mask)));
}
static inline zig_u128 zig_min_u128(zig_u128 lhs, zig_u128 rhs) {
return zig_cmp_u128(lhs, rhs) < INT32_C(0) ? lhs : rhs;
}

59
src/codegen/llvm.zig
View File

@ -7215,20 +7215,28 @@ pub const FuncGen = struct {
return self.buildFloatOp(.floor, inst_ty, 1, .{result});
}
if (scalar_ty.isSignedInt()) {
// const d = @divTrunc(a, b);
// const r = @rem(a, b);
// return if (r == 0) d else d - ((a < 0) ^ (b < 0));
const result_llvm_ty = try self.dg.lowerType(inst_ty);
const zero = result_llvm_ty.constNull();
const div_trunc = self.builder.buildSDiv(lhs, rhs, "");
const target = self.dg.module.getTarget();
const inst_llvm_ty = try self.dg.lowerType(inst_ty);
const scalar_bit_size_minus_one = scalar_ty.bitSize(target) - 1;
const bit_size_minus_one = if (inst_ty.zigTypeTag() == .Vector) const_vector: {
const vec_len = inst_ty.vectorLen();
const scalar_llvm_ty = try self.dg.lowerType(scalar_ty);
const shifts = try self.gpa.alloc(*llvm.Value, vec_len);
defer self.gpa.free(shifts);
@memset(shifts, scalar_llvm_ty.constInt(scalar_bit_size_minus_one, .False));
break :const_vector llvm.constVector(shifts.ptr, vec_len);
} else inst_llvm_ty.constInt(scalar_bit_size_minus_one, .False);
const div = self.builder.buildSDiv(lhs, rhs, "");
const rem = self.builder.buildSRem(lhs, rhs, "");
const rem_eq_0 = self.builder.buildICmp(.EQ, rem, zero, "");
const a_lt_0 = self.builder.buildICmp(.SLT, lhs, zero, "");
const b_lt_0 = self.builder.buildICmp(.SLT, rhs, zero, "");
const a_b_xor = self.builder.buildXor(a_lt_0, b_lt_0, "");
const a_b_xor_ext = self.builder.buildZExt(a_b_xor, div_trunc.typeOf(), "");
const d_sub_xor = self.builder.buildSub(div_trunc, a_b_xor_ext, "");
return self.builder.buildSelect(rem_eq_0, div_trunc, d_sub_xor, "");
const div_sign = self.builder.buildXor(lhs, rhs, "");
const div_sign_mask = self.builder.buildAShr(div_sign, bit_size_minus_one, "");
const zero = inst_llvm_ty.constNull();
const rem_nonzero = self.builder.buildICmp(.NE, rem, zero, "");
const correction = self.builder.buildSelect(rem_nonzero, div_sign_mask, zero, "");
return self.builder.buildNSWAdd(div, correction, "");
}
return self.builder.buildUDiv(lhs, rhs, "");
}
@ -7280,12 +7288,27 @@ pub const FuncGen = struct {
return self.builder.buildSelect(ltz, c, a, "");
}
if (scalar_ty.isSignedInt()) {
const a = self.builder.buildSRem(lhs, rhs, "");
const b = self.builder.buildNSWAdd(a, rhs, "");
const c = self.builder.buildSRem(b, rhs, "");
const target = self.dg.module.getTarget();
const scalar_bit_size_minus_one = scalar_ty.bitSize(target) - 1;
const bit_size_minus_one = if (inst_ty.zigTypeTag() == .Vector) const_vector: {
const vec_len = inst_ty.vectorLen();
const scalar_llvm_ty = try self.dg.lowerType(scalar_ty);
const shifts = try self.gpa.alloc(*llvm.Value, vec_len);
defer self.gpa.free(shifts);
@memset(shifts, scalar_llvm_ty.constInt(scalar_bit_size_minus_one, .False));
break :const_vector llvm.constVector(shifts.ptr, vec_len);
} else inst_llvm_ty.constInt(scalar_bit_size_minus_one, .False);
const rem = self.builder.buildSRem(lhs, rhs, "");
const div_sign = self.builder.buildXor(lhs, rhs, "");
const div_sign_mask = self.builder.buildAShr(div_sign, bit_size_minus_one, "");
const rhs_masked = self.builder.buildAnd(rhs, div_sign_mask, "");
const zero = inst_llvm_ty.constNull();
const ltz = self.builder.buildICmp(.SLT, lhs, zero, "");
return self.builder.buildSelect(ltz, c, a, "");
const rem_nonzero = self.builder.buildICmp(.NE, rem, zero, "");
const correction = self.builder.buildSelect(rem_nonzero, rhs_masked, zero, "");
return self.builder.buildNSWAdd(rem, correction, "");
}
return self.builder.buildURem(lhs, rhs, "");
}

3
test/behavior/math.zig
View File

@ -449,6 +449,9 @@ fn testDivision() !void {
try expect(mod(i32, 10, 12) == 10);
try expect(mod(i32, -14, 12) == 10);
try expect(mod(i32, -2, 12) == 10);
try expect(mod(i32, 10, -12) == -2);
try expect(mod(i32, -14, -12) == -2);
try expect(mod(i32, -2, -12) == -2);
comptime {
try expect(