MerryMage
|
132c783320
|
IR: Implement VectorNarrow
|
2020-04-22 20:46:13 +01:00 |
|
MerryMage
|
1423584f9f
|
constant_pool: Allow for 128-bit constants
|
2020-04-22 20:46:13 +01:00 |
|
MerryMage
|
69de50a878
|
emit_x64_vector: Add SSE4.1 implementations for VectorZeroExtend
|
2020-04-22 20:46:13 +01:00 |
|
MerryMage
|
cbc9f361b0
|
IR: Implement VectorSub
|
2020-04-22 20:46:13 +01:00 |
|
MerryMage
|
b22c5961f9
|
IR: Implement VectorLogicalShiftRight
|
2020-04-22 20:46:13 +01:00 |
|
MerryMage
|
59ace60b03
|
IR: Implement VectorZeroExtend
|
2020-04-22 20:46:13 +01:00 |
|
MerryMage
|
f6247125c0
|
IR: Implement VectorLogicalShiftLeft{8,16,32,64}
|
2020-04-22 20:46:13 +01:00 |
|
MerryMage
|
15e8231f24
|
opcodes: Sort vector IR opcodes alphabetically
|
2020-04-22 20:46:13 +01:00 |
|
MerryMage
|
d74f4e35f6
|
block_of_code: Increase constant pool size
|
2020-04-22 20:46:13 +01:00 |
|
MerryMage
|
e69288f803
|
devirtualize: MinGW uses Intanium MFP ABI
|
2020-04-22 20:46:13 +01:00 |
|
MerryMage
|
ad428cbd7a
|
callback: Properly handle calls with return pointers and simplify interface
|
2020-04-22 20:46:13 +01:00 |
|
MerryMage
|
7a87e3fc55
|
devirtualize: Handle Windows ABI
|
2020-04-22 20:46:13 +01:00 |
|
MerryMage
|
f808a0fbde
|
devirtualize: Devirtualize Itanium ABI MFPs at runtime
|
2020-04-22 20:46:13 +01:00 |
|
Lioncash
|
35a29a9665
|
A64: Implement ZIP1
|
2020-04-22 20:46:13 +01:00 |
|
FernandoS27
|
586854117b
|
Implemented UMULH and SMULH instructions
|
2020-04-22 20:46:13 +01:00 |
|
MerryMage
|
44c3c2312a
|
a64_jitstate: Remove unnecessary FPSCR_nzcv member
|
2020-04-22 20:46:13 +01:00 |
|
MerryMage
|
aac5af50e2
|
IR: FPCompare{32,64} now return NZCV flags instead of implicitly setting them
|
2020-04-22 20:46:13 +01:00 |
|
MerryMage
|
dd2a6684fe
|
IR: Add ConditionalSelectNZCV instruction
|
2020-04-22 20:46:13 +01:00 |
|
MerryMage
|
b173fcf34e
|
backend_x64: Simplify FPDoubleToU32 and FPSingleToU32
They're inaccurate in terms of FPSR at the moment anyway.
|
2020-04-22 20:46:13 +01:00 |
|
Lioncash
|
d040920727
|
Common: Put AES code within its own nested namespace
Prevents the functions from potentially clashing with other stuff in Common in the future
|
2020-04-22 20:46:13 +01:00 |
|
Lioncash
|
40614202e7
|
A64: Implement AESD
|
2020-04-22 20:46:13 +01:00 |
|
Lioncash
|
ccef85dbb7
|
A64: Implement AESE
|
2020-04-22 20:46:13 +01:00 |
|
MerryMage
|
68f46c8334
|
backend_x64: Use a reference to BlockOfCode instead of a pointer
|
2020-04-22 20:46:13 +01:00 |
|
MerryMage
|
8931ee346b
|
IR: Add IR instruction NZCVFromPackedFlags
This instruction expects NZCV to be in the high bits.
i.e.: The positions they were in PSTATE.
|
2020-04-22 20:46:13 +01:00 |
|
MerryMage
|
75b8a76630
|
a64_jitstate: A64 does not have a seperate FPSCR.NZCV
|
2020-04-22 20:46:13 +01:00 |
|
MerryMage
|
6414736a8d
|
emit_x64_vector: bug: VectorGetElement8 returning incorrect values for non-SSE4.1
This bug wasn't discovered earlier because we previously only used index == 0.
|
2020-04-22 20:46:13 +01:00 |
|
MerryMage
|
ebfc51c609
|
IR: Implement VectorSetElement{8,16,32,64}
|
2020-04-22 20:46:13 +01:00 |
|
Lioncash
|
a5c4fbc783
|
A64: Implement AESIMC and AESMC
|
2020-04-22 20:46:13 +01:00 |
|
Lioncash
|
ab9b5fb8aa
|
Common: Relocate common bits of CRC32
Allows the algorithm to be used in any other potential backend.
|
2020-04-22 20:46:12 +01:00 |
|
Lioncash
|
af1384d700
|
A64: Implement CRC32
|
2020-04-22 20:46:12 +01:00 |
|
MerryMage
|
64761dbc72
|
scope_exit: Add SCOPE_SUCCESS and SCOPE_EXIT
|
2020-04-22 20:46:12 +01:00 |
|
MerryMage
|
bafb39ebc5
|
A64: Add Disassemble method
|
2020-04-22 20:46:12 +01:00 |
|
MerryMage
|
f023bbb893
|
A32: Add ExceptionRaised IR instruction and use it
|
2020-04-22 20:46:12 +01:00 |
|
Lioncash
|
7ffbebf290
|
A64: Implement CRC32C
|
2020-04-22 20:46:12 +01:00 |
|
MerryMage
|
d7044bc751
|
assert: Use fmt in ASSERT_MSG
|
2020-04-22 20:46:12 +01:00 |
|
MerryMage
|
52268298a8
|
a64_emit_x64: Perform RSB predictions
|
2020-04-22 20:46:12 +01:00 |
|
MerryMage
|
98ec9c5f90
|
A32: Change UserCallbacks to be similar to A64's interface
|
2020-04-22 20:46:12 +01:00 |
|
Lioncash
|
b9ce660113
|
reg_alloc: std::move RegAlloc's function argument
|
2020-04-22 20:46:12 +01:00 |
|
Lioncash
|
ed561d6653
|
General: Add missing override specifiers
|
2020-04-22 20:46:12 +01:00 |
|
MerryMage
|
b2d99eddc6
|
EmitZeroExtendLongToQuad: Do not rely on register allocator to zero extend 64->128
|
2020-04-22 20:46:12 +01:00 |
|
MerryMage
|
54de64f5bf
|
a64_emit_x64: bug: x64 sign-extends 32-bit immediates
|
2020-04-22 20:46:12 +01:00 |
|
MerryMage
|
6fc228f7fd
|
ir_opt: Add A64 Get/Set Elimination Pass
|
2020-04-22 20:46:12 +01:00 |
|
MerryMage
|
af793c2527
|
{a32,a64}_interface: Predict entrypoint
|
2020-04-22 20:46:12 +01:00 |
|
Lioncash
|
7734cf1050
|
A64: Implement EXTR
|
2020-04-22 20:46:12 +01:00 |
|
MerryMage
|
d497464c9f
|
a64_jitstate: Have 128-bit wide spills
|
2020-04-22 20:44:38 +01:00 |
|
MerryMage
|
b513b2ef05
|
IR: Implement IR instructions A64{Get,Set}S
|
2020-04-22 20:44:38 +01:00 |
|
MerryMage
|
16fa2cd8f6
|
a64_emit_x64: Use xword from Xbyak::util
|
2020-04-22 20:44:38 +01:00 |
|
Lioncash
|
67443efb62
|
General: Convert multiple namespace specifiers to nested namespace specifiers where applicable
Makes namespacing a little less noisy
|
2020-04-22 20:44:38 +01:00 |
|
MerryMage
|
d5283e46e8
|
IR: Implement IR instructions VectorEqual{8,16,32,64,128}
|
2020-04-22 20:44:38 +01:00 |
|
MerryMage
|
4ce9c65cfb
|
reg_alloc: Use std::exchange
|
2020-04-22 20:44:38 +01:00 |
|