aboutsummaryrefslogtreecommitdiff
path: root/lib/Target/X86/X86InstrInfo.cpp
AgeCommit message (Collapse)Author
2008-01-12fix a wordo that gordon noticed :)Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45896 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-12Any x86 instruction that reads from an invariant location is invariant.Chris Lattner
This allows us to sink things like: cvtsi2sd 32(%esp), %xmm1 when reading from the argument area, for example. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45895 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-11rename MachineInstr::setInstrDescriptor -> setDescChris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45871 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-11remove xchg and shift-reg-by-1 instructions, which are dead.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45870 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-10Simplify the side effect stuff a bit more and make licm/sinkingChris Lattner
both work right according to the new flags. This removes the TII::isReallySideEffectFree predicate, and adds TII::isInvariantLoad. It removes NeverHasSideEffects+MayHaveSideEffects and adds UnmodeledSideEffects as machine instr flags. Now the clients can decide everything they need. I think isRematerializable can be implemented in terms of the flags we have now, though I will let others tackle that. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45843 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-10verify that the frame index is immutable before remat'ing (still disabled)Chris Lattner
or being side-effect free. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45816 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-09add a testcase Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45768 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-07Operand 1 should be a register. We don't care if it's a preg, vreg, or 0.Bill Wendling
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45699 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-07rename TargetInstrDescriptor -> TargetInstrDesc.Chris Lattner
Make MachineInstr::getDesc return a reference instead of a pointer, since it can never be null. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45695 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-07Move a bunch more accessors from TargetInstrInfo to TargetInstrDescriptorChris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45680 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-07remove MachineOpCode typedef.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45679 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-07Add predicates methods to TargetOperandInfo, and switch all clients Chris Lattner
over to using them, instead of diddling Flags directly. Change the various flags from const variables to enums. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45677 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-07Rename MachineInstr::getInstrDescriptor -> getDesc(), which reflectsChris Lattner
that it is cheap and efficient to get. Move a variety of predicates from TargetInstrInfo into TargetInstrDescriptor, which makes it much easier to query a predicate when you don't have TII around. Now you can use MI->getDesc()->isBranch() instead of going through TII, and this is much more efficient anyway. Not all of the predicates have been moved over yet. Update old code that used MI->getInstrDescriptor()->Flags to use the new predicates in many places. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45674 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-07Move even more functionality from MRegisterInfo into TargetInstrInfo.Owen Anderson
Some day I'll get it all moved over... git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45672 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-05Fix comment.Bill Wendling
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45638 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-05Chris and Evan noticed that this check was compleatly fubared. I wasBill Wendling
checking that there was a from a global instead of a load from the stub for a global, which is the one that's safe to hoist. Consider this program: volatile char G[100]; int B(char *F, int N) { for (; N > 0; --N) F[N] = G[N]; } In static mode, we shouldn't be hoisting the load from G: $ llc -relocation-model=static -o - a.bc -march=x86 -machine-licm LBB1_1: # bb.preheader leal -1(%eax), %edx testl %edx, %edx movl $1, %edx cmovns %eax, %edx xorl %esi, %esi LBB1_2: # bb movb _G(%eax), %bl movb %bl, (%ecx,%eax) git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45626 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-05enable sinking and licm of loads from the argument area. I'd like to enable ↵Chris Lattner
this for remat, but can't due to an RA bug. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45623 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-05simplify some code by using shorter accessors.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45622 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-05revert my previous patch.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45621 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-05factor some code better to avoid redundancy between Chris Lattner
isReallySideEffectFree and isReallyTriviallyReMaterializable. Why is a load from a global considered side-effect-free but not rematable? git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45620 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-04Move some more functionality from MRegisterInfo to TargetInstrInfo.Owen Anderson
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45603 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-02Machine LICM will check that operands are defined outside of the loop. AlsoBill Wendling
check that register isn't 0 before going further. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45498 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-01Move some more instruction creation methods from RegisterInfo into InstrInfo.Owen Anderson
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45484 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-01Fix a bug in my previous patch: refer to the impl not the pure virtual ↵Chris Lattner
version. It's unclear why gcc would ever compile this... git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45476 91177308-0d34-0410-b5e6-96231b3b80d8
2008-01-01Fix a problem where lib/Target/TargetInstrInfo.h would include and useChris Lattner
a header file from libcodegen. This violates a layering order: codegen depends on target, not the other way around. The fix to this is to split TII into two classes, TII and TargetInstrInfoImpl, which defines stuff that depends on libcodegen. It is defined in libcodegen, where the base is not. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45475 91177308-0d34-0410-b5e6-96231b3b80d8
2007-12-31Move copyRegToReg from MRegisterInfo to TargetInstrInfo. This is part of theOwen Anderson
Machine-level API cleanup instigated by Chris. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45470 91177308-0d34-0410-b5e6-96231b3b80d8
2007-12-31Rename SSARegMap -> MachineRegisterInfo in keeping with the idea Chris Lattner
that "machine" classes are used to represent the current state of the code being compiled. Given this expanded name, we can start moving other stuff into it. For now, move the UsedPhysRegs and LiveIn/LoveOuts vectors from MachineFunction into it. Update all the clients to match. This also reduces some needless #includes, such as MachineModuleInfo from MachineFunction. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45467 91177308-0d34-0410-b5e6-96231b3b80d8
2007-12-30Add new shorter predicates for testing machine operands for various types: Chris Lattner
e.g. MO.isMBB() instead of MO.isMachineBasicBlock(). I don't plan on switching everything over, so new clients should just start using the shorter names. Remove old long accessors, switching everything over to use the short accessor: getMachineBasicBlock() -> getMBB(), getConstantPoolIndex() -> getIndex(), setMachineBasicBlock -> setMBB(), etc. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45464 91177308-0d34-0410-b5e6-96231b3b80d8
2007-12-30Use MachineOperand::getImm instead of MachineOperand::getImmedValue. ↵Chris Lattner
Likewise setImmedValue -> setImm git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45453 91177308-0d34-0410-b5e6-96231b3b80d8
2007-12-30If we have a load of a global address that's not modified during theBill Wendling
function, then go ahead and hoist it out of the loop. This is the result: $ cat a.c volatile int G; int A(int N) { for (; N > 0; --N) G++; } $ llc -o - -relocation-model=pic _A: ... LBB1_2: # bb movl L_G$non_lazy_ptr-"L1$pb"(%eax), %esi incl (%esi) incl %edx cmpl %ecx, %edx jne LBB1_2 # bb ... $ llc -o - -relocation-model=pic -machine-licm _A: ... movl L_G$non_lazy_ptr-"L1$pb"(%eax), %eax LBB1_2: # bb incl (%eax) incl %edx cmpl %ecx, %edx jne LBB1_2 # bb ... I'm limiting this to the MOV32rm x86 instruction for now. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45444 91177308-0d34-0410-b5e6-96231b3b80d8
2007-12-29Remove attribution from file headers, per discussion on llvmdev.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45418 91177308-0d34-0410-b5e6-96231b3b80d8
2007-12-17Add "mayHaveSideEffects" and "neverHasSideEffects" flags to some instructions. IBill Wendling
based what flag to set on whether it was already marked as "isRematerializable". If there was a further check to determine if it's "really" rematerializable, then I marked it as "mayHaveSideEffects" and created a check in the X86 back-end similar to the remat one. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@45132 91177308-0d34-0410-b5e6-96231b3b80d8
2007-12-08Reverting 44702. It wasn't correct to rename them.Bill Wendling
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@44727 91177308-0d34-0410-b5e6-96231b3b80d8
2007-12-08Renaming:Bill Wendling
isTriviallyReMaterializable -> hasNoSideEffects isReallyTriviallyReMaterializable -> isTriviallyReMaterializable git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@44702 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-11Added tail call optimization to the x86 back end. It can beArnold Schwaighofer
enabled by passing -tailcallopt to llc. The optimization is performed if the following conditions are satisfied: * caller/callee are fastcc * elf/pic is disabled OR elf/pic enabled + callee is in module + callee has visibility protected or hidden git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42870 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-09Under 64-bit mode use LEA64_32r instead of LEA64r to save a byte.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42783 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-08Allow x86 compare to be commutable by default.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42761 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-05Commute x86 cmove instructions by swapping the operands and change the conditionEvan Cheng
to its inverse. Testing this as llcbeta git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42661 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-05Enable convertToThreeAddress for X86 by default.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42655 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-05INC64_32r -> LEA64_32r is better than INC64_32r -> LEA32r, but it still canEvan Cheng
cause performance degradation. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42653 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-05In 64-bit mode, avoid using leal with 32-bit 32-bit address size, e.g.Evan Cheng
leal 1(%ecx), %edi, which requires 67H prefix. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42647 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-05Add support to convert more 64-bit instructions to 3-address instructions.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42642 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-05Testing convertToThreeeAddress as X86 llcbeta.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42630 91177308-0d34-0410-b5e6-96231b3b80d8
2007-09-29Enabling new condition code modeling scheme.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42459 91177308-0d34-0410-b5e6-96231b3b80d8
2007-09-25Added support for new condition code modeling scheme (i.e. physical register ↵Evan Cheng
dependency). These are a bunch of instructions that are duplicated so the x86 backend can support both the old and new schemes at the same time. They will be deleted after all the kinks are worked out. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42285 91177308-0d34-0410-b5e6-96231b3b80d8
2007-09-17Add 64-bit jmp instructions to the list of instructions thatDan Gohman
can terminate a block with no fall-through. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42029 91177308-0d34-0410-b5e6-96231b3b80d8
2007-09-14Add patterns for SHLD64* and SHRD64*.Dan Gohman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@41975 91177308-0d34-0410-b5e6-96231b3b80d8
2007-09-14Add implicit def of EFLAGS on those instructions that may modify flags.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@41962 91177308-0d34-0410-b5e6-96231b3b80d8
2007-09-07Add lengthof and endof templates that hide a lot of sizeof computations.Owen Anderson
Patch by Sterling Stein! git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@41758 91177308-0d34-0410-b5e6-96231b3b80d8
2007-09-06Fix a bug in X86InstrInfo::convertToThreeAddress that caused it to codegen:Evan Cheng
leal (,%rcx,8), %rcx It should be leal (,%rcx,8), %ecx git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@41735 91177308-0d34-0410-b5e6-96231b3b80d8
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2025-09-01 01:48:41 +0000