RGBCube/serenity
1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-06-01 09:58:14 +00:00
Code Issues Activity Actions

UserspaceEmulator: Simplify the definition of the FPU register stack

Browse source 
Long doubles are always at least 80 bits wide in memory and it suffices
if we can address these 80 bits, to mark the long double as NAN at the
end of an MMX instruction, so the additional magic using conditional
types is unnecessary.
This commit is contained in:
Hendiadyoin1 2021-07-23 00:40:28 +02:00 • committed by Linus Groh
parent d06675e3e4
commit 7214b08f81
2 changed files with 3 additions and 9 deletions
Download patch file Download diff file Expand all files Collapse all files

5
Userland/DevTools/UserspaceEmulator/SoftFPU.cpp
View file

@ -93,11 +93,10 @@ ALWAYS_INLINE MMX SoftFPU::mmx_get(u8 index) const
ALWAYS_INLINE void SoftFPU::mmx_set(u8 index, MMX value)
{
m_storage[index].mmx = value;
// The high bytes are set to 0b11... to make the floatingpoint value NaN.
// The high bytes are set to 0b11... to make the floating-point value NaN.
// This way we are technically able to find out if we are reading the wrong
// type, but this is still difficult, so we use our own lookup for that
// We set the alignment bytes to all 1's, too, just in case
m_storage[index].__high = ~(decltype(m_storage[index].__high))0u;
m_storage[index].__high = 0xFFFFU;
m_reg_is_mmx[index] = true;
}