[libc-commits] [libc] [libc][NFC] fix int warnings in float conversion (PR #74379)
via libc-commits
libc-commits at lists.llvm.org
Mon Dec 4 16:04:31 PST 2023
https://github.com/michaelrj-google updated https://github.com/llvm/llvm-project/pull/74379
>From 47091bc1052edc7b795f73320f5c5fe163a2b972 Mon Sep 17 00:00:00 2001
From: Michael Jones <michaelrj at google.com>
Date: Mon, 4 Dec 2023 14:17:04 -0800
Subject: [PATCH 1/3] [libc][NFC] fix int warnings in float conversion
The printf float to string conversion functions had some implicit
integer conversion warnings on gcc. This patch adds explicit casts to
these places.
---
libc/src/__support/UInt.h | 1 +
libc/src/__support/float_to_string.h | 19 +++++++++++--------
libc/src/math/generic/math_utils.h | 4 +++-
.../stdio/printf_core/float_dec_converter.h | 13 ++++++++-----
.../stdio/printf_core/float_hex_converter.h | 6 +++---
5 files changed, 26 insertions(+), 17 deletions(-)
diff --git a/libc/src/__support/UInt.h b/libc/src/__support/UInt.h
index dbedfbc4c197e..248324c326dfb 100644
--- a/libc/src/__support/UInt.h
+++ b/libc/src/__support/UInt.h
@@ -121,6 +121,7 @@ template <size_t Bits, bool Signed> struct BigInt {
return lo;
}
} else {
+ // TODO: silence shift warning
return static_cast<T>((static_cast<T>(val[1]) << 64) + lo);
}
}
diff --git a/libc/src/__support/float_to_string.h b/libc/src/__support/float_to_string.h
index e198a5a7215fb..1a3e1385f606b 100644
--- a/libc/src/__support/float_to_string.h
+++ b/libc/src/__support/float_to_string.h
@@ -136,12 +136,12 @@ LIBC_INLINE constexpr uint32_t log10_pow2(const uint64_t e) {
// us the floor, whereas counting the digits of the power of 2 gives us the
// ceiling. With a similar loop I checked the maximum valid value and found
// 42039.
- return (e * 0x13441350fbdll) >> 42;
+ return static_cast<uint32_t>((e * 0x13441350fbdll) >> 42);
}
// Same as above, but with different constants.
LIBC_INLINE constexpr uint32_t log2_pow5(const uint64_t e) {
- return (e * 0x12934f0979bll) >> 39;
+ return static_cast<uint32_t>((e * 0x12934f0979bll) >> 39);
}
// Returns 1 + floor(log_10(2^e). This could technically be off by 1 if any
@@ -157,9 +157,10 @@ LIBC_INLINE constexpr uint32_t ceil_log10_pow2(const uint32_t e) {
LIBC_INLINE constexpr uint32_t length_for_num(const uint32_t idx,
const uint32_t mantissa_width) {
//+8 to round up when dividing by 9
- return (ceil_log10_pow2(idx) + ceil_log10_pow2(mantissa_width + 1) +
- (BLOCK_SIZE - 1)) /
- BLOCK_SIZE;
+ return static_cast<uint32_t>((ceil_log10_pow2(idx) +
+ ceil_log10_pow2(mantissa_width + 1) +
+ (BLOCK_SIZE - 1)) /
+ BLOCK_SIZE);
// return (ceil_log10_pow2(16 * idx + mantissa_width) + 8) / 9;
}
@@ -488,7 +489,8 @@ class FloatToString {
val = POW10_SPLIT[POW10_OFFSET[idx] + block_index];
#endif
- const uint32_t shift_amount = SHIFT_CONST + (IDX_SIZE * idx) - exponent;
+ const uint32_t shift_amount =
+ static_cast<uint32_t>(SHIFT_CONST + (IDX_SIZE * idx) - exponent);
const uint32_t digits =
internal::mul_shift_mod_1e9(mantissa, val, (int32_t)(shift_amount));
return digits;
@@ -549,7 +551,7 @@ class FloatToString {
val = POW10_SPLIT_2[p];
#endif
const int32_t shift_amount =
- SHIFT_CONST + (-exponent - static_cast<int>(IDX_SIZE) * idx);
+ static_cast<int32_t>(SHIFT_CONST + (-exponent - IDX_SIZE * idx));
uint32_t digits =
internal::mul_shift_mod_1e9(mantissa, val, shift_amount);
return digits;
@@ -747,7 +749,8 @@ FloatToString<long double>::get_negative_block(int block_index) {
block_index + 1);
}
#endif
- const int32_t shift_amount = SHIFT_CONST + (-exponent - IDX_SIZE * idx);
+ const int32_t shift_amount =
+ SHIFT_CONST + (-exponent - static_cast<int>(IDX_SIZE * idx));
BlockInt digits = internal::mul_shift_mod_1e9(mantissa, val, shift_amount);
return digits;
} else {
diff --git a/libc/src/math/generic/math_utils.h b/libc/src/math/generic/math_utils.h
index 38a14a47e88fa..e7fb193a08bea 100644
--- a/libc/src/math/generic/math_utils.h
+++ b/libc/src/math/generic/math_utils.h
@@ -34,7 +34,9 @@ LIBC_INLINE double as_double(uint64_t x) { return cpp::bit_cast<double>(x); }
LIBC_INLINE uint32_t top12_bits(float x) { return as_uint32_bits(x) >> 20; }
-LIBC_INLINE uint32_t top12_bits(double x) { return as_uint64_bits(x) >> 52; }
+LIBC_INLINE uint32_t top12_bits(double x) {
+ return static_cast<uint32_t>(as_uint64_bits(x) >> 52);
+}
// Values to trigger underflow and overflow.
template <typename T> struct XFlowValues;
diff --git a/libc/src/stdio/printf_core/float_dec_converter.h b/libc/src/stdio/printf_core/float_dec_converter.h
index 0e152a2602564..f329ba6e1dd3e 100644
--- a/libc/src/stdio/printf_core/float_dec_converter.h
+++ b/libc/src/stdio/printf_core/float_dec_converter.h
@@ -537,7 +537,7 @@ LIBC_INLINE int convert_float_decimal_typed(Writer *writer,
}
if (exponent < MANT_WIDTH) {
- const uint32_t blocks = (precision / BLOCK_SIZE) + 1;
+ const uint32_t blocks = static_cast<uint32_t>(precision / BLOCK_SIZE) + 1;
uint32_t i = 0;
// if all the blocks we should write are zero
if (blocks <= float_converter.zero_blocks_after_point()) {
@@ -561,7 +561,8 @@ LIBC_INLINE int convert_float_decimal_typed(Writer *writer,
RET_IF_RESULT_NEGATIVE(float_writer.write_middle_block(digits));
} else {
- const uint32_t maximum = precision - BLOCK_SIZE * i;
+ const uint32_t maximum =
+ static_cast<uint32_t>(precision - BLOCK_SIZE * i);
uint32_t last_digit = 0;
for (uint32_t k = 0; k < BLOCK_SIZE - maximum; ++k) {
last_digit = digits % 10;
@@ -646,7 +647,8 @@ LIBC_INLINE int convert_float_dec_exp_typed(Writer *writer,
const size_t block_width = IntegerToString<intmax_t>(digits).size();
- final_exponent = (cur_block * BLOCK_SIZE) + static_cast<int>(block_width - 1);
+ final_exponent = static_cast<int>(cur_block * BLOCK_SIZE) +
+ static_cast<int>(block_width - 1);
int positive_exponent = final_exponent < 0 ? -final_exponent : final_exponent;
size_t exponent_width = IntegerToString<intmax_t>(positive_exponent).size();
@@ -819,7 +821,8 @@ LIBC_INLINE int convert_float_dec_auto_typed(Writer *writer,
size_t trailing_zeroes = 0;
size_t trailing_nines = 0;
- base_10_exp = (cur_block * BLOCK_SIZE) + static_cast<int>(block_width - 1);
+ base_10_exp = static_cast<int>(cur_block * BLOCK_SIZE) +
+ static_cast<int>(block_width - 1);
// If the first block is not also the last block
if (block_width <= exp_precision + 1) {
@@ -864,7 +867,7 @@ LIBC_INLINE int convert_float_dec_auto_typed(Writer *writer,
(cur_last_digit == 9 || cur_last_digit == 0)) {
// If the next digit is not the same as the previous one, then there are
// no more contiguous trailing digits.
- if ((copy_of_digits % 10) != cur_last_digit) {
+ if (static_cast<int>(copy_of_digits % 10) != cur_last_digit) {
break;
}
if (cur_last_digit == 9) {
diff --git a/libc/src/stdio/printf_core/float_hex_converter.h b/libc/src/stdio/printf_core/float_hex_converter.h
index 6a980a74d4a6f..1f105492e8e5a 100644
--- a/libc/src/stdio/printf_core/float_hex_converter.h
+++ b/libc/src/stdio/printf_core/float_hex_converter.h
@@ -139,8 +139,8 @@ LIBC_INLINE int convert_float_hex_exp(Writer *writer,
size_t first_non_zero = 1;
for (; mant_cur > 0; --mant_cur, mantissa >>= 4) {
char mant_mod_16 = static_cast<char>(mantissa) & 15;
- char new_digit =
- (mant_mod_16 > 9) ? (mant_mod_16 - 10 + a) : (mant_mod_16 + '0');
+ char new_digit = static_cast<char>(
+ (mant_mod_16 > 9) ? (mant_mod_16 - 10 + a) : (mant_mod_16 + '0'));
mant_buffer[mant_cur - 1] = new_digit;
if (new_digit != '0' && first_non_zero < mant_cur)
first_non_zero = mant_cur;
@@ -169,7 +169,7 @@ LIBC_INLINE int convert_float_hex_exp(Writer *writer,
size_t exp_cur = EXP_LEN;
for (; exponent > 0; --exp_cur, exponent /= 10) {
- exp_buffer[exp_cur - 1] = (exponent % 10) + '0';
+ exp_buffer[exp_cur - 1] = static_cast<char>((exponent % 10) + '0');
}
if (exp_cur == EXP_LEN) { // if nothing else was written, write a 0.
exp_buffer[EXP_LEN - 1] = '0';
>From 170b4b558a90fbc7e272ba6bd15b2057783271aa Mon Sep 17 00:00:00 2001
From: Michael Jones <michaelrj at google.com>
Date: Mon, 4 Dec 2023 14:33:26 -0800
Subject: [PATCH 2/3] Rebase and narrow scope of casts in float_to_string
Patch #74369 fixed one of the fixed in the previous commit. This rebases
on top of that commit and also adjusts the casts in that file to better
match.
---
libc/src/__support/float_to_string.h | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/libc/src/__support/float_to_string.h b/libc/src/__support/float_to_string.h
index 1a3e1385f606b..c93c87ec53f87 100644
--- a/libc/src/__support/float_to_string.h
+++ b/libc/src/__support/float_to_string.h
@@ -287,11 +287,11 @@ LIBC_INLINE cpp::UInt<MID_INT_SIZE> get_table_negative(int exponent, size_t i) {
if (block_shifts < static_cast<int>(ten_blocks)) {
ten_blocks = ten_blocks - block_shifts;
five_blocks = block_shifts;
- shift_amount = shift_amount + (block_shifts * BLOCK_SIZE);
+ shift_amount = shift_amount + static_cast<int>(block_shifts * BLOCK_SIZE);
} else {
ten_blocks = 0;
five_blocks = i;
- shift_amount = static_cast<int>(shift_amount + (i * BLOCK_SIZE));
+ shift_amount = shift_amount + static_cast<int>(i * BLOCK_SIZE);
}
}
@@ -490,7 +490,7 @@ class FloatToString {
val = POW10_SPLIT[POW10_OFFSET[idx] + block_index];
#endif
const uint32_t shift_amount =
- static_cast<uint32_t>(SHIFT_CONST + (IDX_SIZE * idx) - exponent);
+ SHIFT_CONST + static_cast<uint32_t>(IDX_SIZE * idx) - exponent;
const uint32_t digits =
internal::mul_shift_mod_1e9(mantissa, val, (int32_t)(shift_amount));
return digits;
@@ -551,7 +551,7 @@ class FloatToString {
val = POW10_SPLIT_2[p];
#endif
const int32_t shift_amount =
- static_cast<int32_t>(SHIFT_CONST + (-exponent - IDX_SIZE * idx));
+ SHIFT_CONST + (-exponent - static_cast<int32_t>(IDX_SIZE * idx));
uint32_t digits =
internal::mul_shift_mod_1e9(mantissa, val, shift_amount);
return digits;
>From dec336770ee8db6450fa16970691ae703a58ff3d Mon Sep 17 00:00:00 2001
From: Michael Jones <michaelrj at google.com>
Date: Mon, 4 Dec 2023 16:04:03 -0800
Subject: [PATCH 3/3] more type tweaks
---
libc/src/__support/float_to_string.h | 17 ++++++++---------
libc/src/math/generic/math_utils.h | 19 +++----------------
.../stdio/printf_core/float_dec_converter.h | 4 ++--
3 files changed, 13 insertions(+), 27 deletions(-)
diff --git a/libc/src/__support/float_to_string.h b/libc/src/__support/float_to_string.h
index c93c87ec53f87..34c0c0ceef286 100644
--- a/libc/src/__support/float_to_string.h
+++ b/libc/src/__support/float_to_string.h
@@ -103,7 +103,7 @@ constexpr size_t MID_INT_SIZE = 192;
namespace LIBC_NAMESPACE {
using BlockInt = uint32_t;
-constexpr size_t BLOCK_SIZE = 9;
+constexpr uint32_t BLOCK_SIZE = 9;
using MantissaInt = fputil::FPBits<long double>::UIntType;
@@ -157,10 +157,9 @@ LIBC_INLINE constexpr uint32_t ceil_log10_pow2(const uint32_t e) {
LIBC_INLINE constexpr uint32_t length_for_num(const uint32_t idx,
const uint32_t mantissa_width) {
//+8 to round up when dividing by 9
- return static_cast<uint32_t>((ceil_log10_pow2(idx) +
- ceil_log10_pow2(mantissa_width + 1) +
- (BLOCK_SIZE - 1)) /
- BLOCK_SIZE);
+ return (ceil_log10_pow2(idx) + ceil_log10_pow2(mantissa_width + 1) +
+ (BLOCK_SIZE - 1)) /
+ BLOCK_SIZE;
// return (ceil_log10_pow2(16 * idx + mantissa_width) + 8) / 9;
}
@@ -287,11 +286,11 @@ LIBC_INLINE cpp::UInt<MID_INT_SIZE> get_table_negative(int exponent, size_t i) {
if (block_shifts < static_cast<int>(ten_blocks)) {
ten_blocks = ten_blocks - block_shifts;
five_blocks = block_shifts;
- shift_amount = shift_amount + static_cast<int>(block_shifts * BLOCK_SIZE);
+ shift_amount = shift_amount + (block_shifts * BLOCK_SIZE);
} else {
ten_blocks = 0;
five_blocks = i;
- shift_amount = shift_amount + static_cast<int>(i * BLOCK_SIZE);
+ shift_amount = shift_amount + (static_cast<int>(i) * BLOCK_SIZE);
}
}
@@ -490,7 +489,7 @@ class FloatToString {
val = POW10_SPLIT[POW10_OFFSET[idx] + block_index];
#endif
const uint32_t shift_amount =
- SHIFT_CONST + static_cast<uint32_t>(IDX_SIZE * idx) - exponent;
+ SHIFT_CONST + (static_cast<uint32_t>(IDX_SIZE) * idx) - exponent;
const uint32_t digits =
internal::mul_shift_mod_1e9(mantissa, val, (int32_t)(shift_amount));
return digits;
@@ -551,7 +550,7 @@ class FloatToString {
val = POW10_SPLIT_2[p];
#endif
const int32_t shift_amount =
- SHIFT_CONST + (-exponent - static_cast<int32_t>(IDX_SIZE * idx));
+ SHIFT_CONST + (-exponent - (static_cast<int32_t>(IDX_SIZE) * idx));
uint32_t digits =
internal::mul_shift_mod_1e9(mantissa, val, shift_amount);
return digits;
diff --git a/libc/src/math/generic/math_utils.h b/libc/src/math/generic/math_utils.h
index e7fb193a08bea..e884fe2deae28 100644
--- a/libc/src/math/generic/math_utils.h
+++ b/libc/src/math/generic/math_utils.h
@@ -18,26 +18,13 @@
#include <stdint.h>
-namespace LIBC_NAMESPACE {
-
-LIBC_INLINE uint32_t as_uint32_bits(float x) {
- return cpp::bit_cast<uint32_t>(x);
-}
-
-LIBC_INLINE uint64_t as_uint64_bits(double x) {
- return cpp::bit_cast<uint64_t>(x);
-}
+// TODO: evaluate which functions from this file are actually used.
-LIBC_INLINE float as_float(uint32_t x) { return cpp::bit_cast<float>(x); }
+namespace LIBC_NAMESPACE {
+// TODO: Remove this, or move it to exp_utils.cpp which is its only user.
LIBC_INLINE double as_double(uint64_t x) { return cpp::bit_cast<double>(x); }
-LIBC_INLINE uint32_t top12_bits(float x) { return as_uint32_bits(x) >> 20; }
-
-LIBC_INLINE uint32_t top12_bits(double x) {
- return static_cast<uint32_t>(as_uint64_bits(x) >> 52);
-}
-
// Values to trigger underflow and overflow.
template <typename T> struct XFlowValues;
diff --git a/libc/src/stdio/printf_core/float_dec_converter.h b/libc/src/stdio/printf_core/float_dec_converter.h
index f329ba6e1dd3e..d1b6633504cb2 100644
--- a/libc/src/stdio/printf_core/float_dec_converter.h
+++ b/libc/src/stdio/printf_core/float_dec_converter.h
@@ -861,13 +861,13 @@ LIBC_INLINE int convert_float_dec_auto_typed(Writer *writer,
trailing_nines = 0;
trailing_zeroes = 0;
BlockInt copy_of_digits = digits;
- int cur_last_digit = copy_of_digits % 10;
+ BlockInt cur_last_digit = copy_of_digits % 10;
// We only care if it ends in nines or zeroes.
while (copy_of_digits > 0 &&
(cur_last_digit == 9 || cur_last_digit == 0)) {
// If the next digit is not the same as the previous one, then there are
// no more contiguous trailing digits.
- if (static_cast<int>(copy_of_digits % 10) != cur_last_digit) {
+ if (copy_of_digits % 10 != cur_last_digit) {
break;
}
if (cur_last_digit == 9) {
More information about the libc-commits
mailing list