- 07b7ea1 Remove -always-fold-and-in-test. by Evan Cheng · 17 years ago
- e526d8a Set to default: x86 no longer fold and into test if it has more than one use. by Evan Cheng · 17 years ago
- 5459143 Revert the assert for MUL_LOHI with an unused high result; Chris by Dan Gohman · 17 years ago
- c2d9b5f Add an assert to verify that we don't see an by Dan Gohman · 17 years ago
- fcf81db Remove the hack that turned an {S,U}MUL_LOHI with an unused high by Dan Gohman · 17 years ago
- 2e68b6f Convert MaskedValueIsZero and all its users to use APInt. Also add by Dan Gohman · 17 years ago
- 3738f2d Poorly named option. by Evan Cheng · 17 years ago
- 79964fd Disable for now. This is pessimizing code. by Evan Cheng · 17 years ago
- e9c608d Add hidden option -x86-fold-and-in-test to test the effect the test / and folding change. by Evan Cheng · 17 years ago
- 9f143ce Only using x86-64 rip relative addressing in non-staic mode? by Evan Cheng · 18 years ago
- 6f0d024 Rename MRegisterInfo to TargetRegisterInfo. by Dan Gohman · 18 years ago
- be3bf42 Fix a x86-64 codegen deficiency. Allow gv + offset when using rip addressing mode. by Evan Cheng · 18 years ago
- 4e3f5a4 Dwarf requires variable entries to be in the source order. Right now, since we are recording variable information at isel time this means parameters would appear in the reverse order. The short term fix is to issue recordVariable() at asm printing time instead. by Evan Cheng · 18 years ago
- a844bde SDIsel processes llvm.dbg.declare by recording the variable debug information descriptor and its corresponding stack frame index in MachineModuleInfo. This only works if the local variable is "homed" in the stack frame. It does not work for byval parameter, etc. by Evan Cheng · 18 years ago
- 0d9e976 Work in progress. This patch *fixes* x86-64 calls which are modelled as StructRet but really should be return in registers, e.g. _Complex long double, some 128-bit aggregates. This is a short term solution that is necessary only because llvm, for now, cannot model i128 nor call's with multiple results. by Evan Cheng · 18 years ago
- d43d00c Significantly simplify and improve handling of FP function results on x86-32. by Chris Lattner · 18 years ago
- b1a9aec Fix a x86-64 static codegen bug. This fixes a lot of x86-64 jit failures. by Evan Cheng · 18 years ago
- 0475ab5 Combine MovePCtoStack + POP32r into one instruction MOVPC32r so it can be moved if needed. by Evan Cheng · 18 years ago
- 84bc542 Rename SSARegMap -> MachineRegisterInfo in keeping with the idea by Chris Lattner · 18 years ago
- 4ee451d Remove attribution from file headers, per discussion on llvmdev. by Chris Lattner · 18 years ago
- f02ca69 Fix JIT code emission of X86::MovePCtoStack. by Evan Cheng · 18 years ago
- 1314b00 Fold some and + shift in x86 addressing mode. by Evan Cheng · 18 years ago
- 5aaddaa aesthetic changes, no functionality change. Evan, it's not clear by Chris Lattner · 18 years ago
- 8a59448 Fix a long standing deficiency in the X86 backend: we would by Chris Lattner · 18 years ago
- 0d64287 Silence, accersed warning by Bill Wendling · 18 years ago
- 8368805 Fix the folding of multiplication into addresses on x86, which was broken by Dan Gohman · 18 years ago
- 96aaa54 Flag MOV32to32_ with EXTRACT_SUBREG. They should not be scheduled apart. by Evan Cheng · 18 years ago
- 74f87a6 Fix grammar in a comment. by Dan Gohman · 18 years ago
- 525178c Migrate X86 and ARM from using X86ISD::{,I}DIV and ARMISD::MULHILO{U,S} to by Dan Gohman · 18 years ago
- 2fe1259 Partly revert invalid r41774 by Anton Korobeynikov · 18 years ago
- a37c9f7 When both x/y and x%y are needed (x and y both scalar integer), compute by Dan Gohman · 18 years ago
- 48d1e45 When mixing SSE and x87 codegen, it's possible to by Dale Johannesen · 18 years ago
- ef61ed3 TableGen no longer emit CopyFromReg nodes for implicit results in physical by Evan Cheng · 18 years ago
- cdd509a Apply feedback from previous patch. by Dale Johannesen · 18 years ago
- eaf0894 Enhance APFloat to retain bits of NaNs (fixes oggenc). by Dale Johannesen · 18 years ago
- badb2d2 When x86 addresses matching exceeds its recursion limit, check to by Dan Gohman · 18 years ago
- a1eb155 Use subregs to improve any_extend code generation when feasible. by Christopher Lamb · 18 years ago
- c59e521 Increase efficiency of sign_extend_inreg by using subregisters for truncation. As the README suggests sign_extend_subreg is selected to (sext(trunc)). by Christopher Lamb · 18 years ago
- f7ef26e divb / mulb outputs to ah. Under x86-64 it's not legal to read ah if the instruction requires a rex prefix (i.e. outputs to r8b, etc.). So issue shift right by 8 on AX and then truncate it to 8 bits instead. by Evan Cheng · 18 years ago
- cdbe4d3 Long double patch 8 of N: make it partially work in by Dale Johannesen · 18 years ago
- 6a30811 Get X86 long double calling convention to work by Dale Johannesen · 18 years ago
- 7afa166 Switch some multiplication instructions over to the new scheme for testing. by Evan Cheng · 18 years ago
- f6844ca Mac OS X X86-64 low 4G address not available. by Evan Cheng · 18 years ago
- 2dc6dc6 Change the x86 backend to use extract_subreg for truncation operations. Passes DejaGnu, SingleSource and MultiSource. by Christopher Lamb · 18 years ago
- cf5543c Minor bug. by Evan Cheng · 18 years ago
- 518143d Same goes for constantpool, etc. by Evan Cheng · 18 years ago
- a49ed78 Mac OS X x86-64 lower 4G address is not available. by Evan Cheng · 18 years ago
- dc9b3d0 Add const to CanBeFoldedBy, CheckAndMask, and CheckOrMask. by Dan Gohman · 18 years ago
- 849f214 Fix for PR 1505 (and 1489). Rewrite X87 register by Dale Johannesen · 18 years ago
- ea859be Move ComputeMaskedBits, MaskedValueIsZero, and ComputeNumSignBits from by Dan Gohman · 18 years ago
- a1b253f Fix CodeGen/X86/2007-03-24-InlineAsmPModifier.ll by Chris Lattner · 18 years ago
- f6e9353 Oops :) by Anton Korobeynikov · 18 years ago
- 33bf8c4 Don't allow MatchAddress recurse too much. This trims exponential by Anton Korobeynikov · 18 years ago
- a16b7cb Two changes: by Chris Lattner · 18 years ago
- 6241226 Fix a miscompilation in the addr mode code trying to implement X | C and by Chris Lattner · 19 years ago
- 706535d Linux GOT indirect reference is only necessary in PIC mode. by Evan Cheng · 19 years ago
- 7aa8a45 Adjust #includes to compensate for lost of DerivedTypes.h in TargetLowering.h by Reid Spencer · 19 years ago
- 7f70559 * PIC codegen for X86/Linux has been implemented by Anton Korobeynikov · 19 years ago
- 317848f Really big cleanup. by Anton Korobeynikov · 19 years ago
- 95b2c7d eliminate static ctors for Statistic objects. by Chris Lattner · 19 years ago
- a70d14b Fix for PR1062 by Dan Gohman. by Evan Cheng · 19 years ago
- f5da133 What should be the last unnecessary <iostream>s in the library. by Bill Wendling · 19 years ago
- ac0b6ae Detemplatize the Statistic class. The only type it is instantiated with by Chris Lattner · 19 years ago
- 953fa04 Revert an unintended change. by Evan Cheng · 19 years ago
- 28b51439 - Switch X86-64 JIT to large code size model. by Evan Cheng · 19 years ago
- 19f2ffc - Fix X86-64 JIT by temporarily disabling code that treats GV address as 32-bit by Evan Cheng · 19 years ago
- 0085a28 - Use a different wrapper node for RIP-relative GV, etc. by Evan Cheng · 19 years ago
- 4946399 Clean up. by Evan Cheng · 19 years ago
- d0ff02c Fix for PR1018 - Better support for X86-64 Linux in small code model. by Evan Cheng · 19 years ago
- c0f64ff Change MachineInstr ctor's to take a TargetInstrDescriptor reference instead by Evan Cheng · 19 years ago
- b1409ce For unsigned 8-bit division. Use movzbw to set the lower 8 bits of AX while by Evan Cheng · 19 years ago
- 6345d75 Removed even more std::cerr and #include <iostream> things. by Bill Wendling · 19 years ago
- 7ce4578 Matches MachineInstr changes. by Evan Cheng · 19 years ago
- 5cd3e9f Add implicit use / def operands to created MI's. by Evan Cheng · 19 years ago
- 490ce1e Add all implicit defs to FP_REG_KILL mi. by Evan Cheng · 19 years ago
- d6373bc Fix a bug in SelectScalarSSELoad. Since the load is wrapped in a by Evan Cheng · 19 years ago
- 0d53826 Match tblegen changes. by Evan Cheng · 19 years ago
- d41b30d Unbreak VC++ build. by Jeff Cohen · 19 years ago
- 1509254 silence warning by Chris Lattner · 19 years ago
- 07e4b00 SelectScalarSSELoad should call CanBeFoldedBy as well. by Evan Cheng · 19 years ago
- 27e1fe9 Corrected load folding check. We need to start from the root of the sub-dag by Evan Cheng · 19 years ago
- 8b2794a Merge ISD::TRUNCSTORE to ISD::STORE. Switch to using StoreSDNode. by Evan Cheng · 19 years ago
- 494cec6 Doh. This wasn't causing problems by luck. by Evan Cheng · 19 years ago
- f78ae9e fix compilation failure of smg2000 by Chris Lattner · 19 years ago
- 4fe4f25 Fold "zero extending vector loads" now that evan added the chain manip stuff. by Chris Lattner · 19 years ago
- 82a9164 ComplexPatterns sse_load_f32 and sse_load_f64 returns in / out chain operands. by Evan Cheng · 19 years ago
- a275ecb More isel time load folding checking for nodes that produce flag values. by Evan Cheng · 19 years ago
- 466685d Reflects ISD::LOAD / ISD::LOADX / LoadSDNode changes. by Evan Cheng · 19 years ago
- 3a7cd95 completely disable folding of loads into scalar sse instructions and provide by Chris Lattner · 19 years ago
- 71f84de Not needed. by Evan Cheng · 19 years ago
- bcb9770 Added some eye-candy for Subtarget type checking by Anton Korobeynikov · 19 years ago
- c45a2c7 Remove a unnecessary check. by Evan Cheng · 19 years ago
- 0f27fc3 Fix a regression in the 32-bit port from the 64-bit port landing. by Chris Lattner · 19 years ago
- c356a57 Reflects MachineConstantPoolEntry changes. by Evan Cheng · 19 years ago
- 25ab690 Committing X86-64 support. by Evan Cheng · 19 years ago
- 1453de5 Oops. Bad typo. Without the check of N1.hasOneUse() bad things can happen. by Evan Cheng · 19 years ago
- 4a88858 Remove dead code. by Evan Cheng · 19 years ago
- 780413d Don't performance load/op/store transformation if op produces a floating point by Evan Cheng · 19 years ago
- e50794a - Enable x86 isel preprocessing by default unless -fast is specified. by Evan Cheng · 19 years ago
- 82a35b3 Avoid making unneeded load/mod/store transformation which can hurt performance. by Evan Cheng · 19 years ago