RGBCube/serenity
1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-07-28 11:37:45 +00:00
Code Issues Activity Actions

AK: Implement FloatExtractor<f128>

Browse source 
This patch adds support for 128-bit floating points in FloatExtractor.

This is required to build SerenityOS on MacOS/aarch64. It might break
building for Raspberry Pi.
This commit is contained in:
Steffen Rusitschka 2022-11-25 18:15:55 +01:00 • committed by Linus Groh
parent 2e806dab07
commit 1aa07d7328
3 changed files with 39 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

18
AK/BuiltinWrappers.h
View file

@ -97,6 +97,24 @@ inline constexpr int count_leading_zeroes(IntType value)
#endif
}
#ifdef __SIZEOF_INT128__
// This is required for math.cpp internal_scalbn
inline constexpr int count_leading_zeroes(unsigned __int128 value)
{
# if defined(AK_COMPILER_CLANG) || defined(AK_COMPILER_GCC)
return (value > __UINT64_MAX__) ? __builtin_clzll(value >> 64) : 64 + __builtin_clzll(value);
# else
unsigned __int128 mask = (unsigned __int128)1 << 127;
int ret = 0;
while ((value & mask) == 0) {
++ret;
mask >>= 1;
}
return ret;
# endif
}
#endif
// The function will return the number of leading zeroes in the type. If
// the given number is zero, this function will return the number of bits
// in the IntType.