aboutsummaryrefslogtreecommitdiff
path: root/lib/CodeGen/LiveIntervalAnalysis.cpp
AgeCommit message (Collapse)Author
2010-06-18TwoAddressInstructionPass::CoalesceExtSubRegs can insert INSERT_SUBREGJakob Stoklund Olesen
instructions, but it doesn't really understand live ranges, so the first INSERT_SUBREG uses an implicitly defined register. Fix it in LiveVariableAnalysis by adding the <undef> flag. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106333 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-16Allow a register to be redefined multiple times in a basic block.Jakob Stoklund Olesen
LiveVariableAnalysis was a bit picky about a register only being redefined once, but that really isn't necessary. Here is an example of chained INSERT_SUBREGs that we can handle now: 68 %reg1040<def> = INSERT_SUBREG %reg1040, %reg1028<kill>, 14 register: %reg1040 +[70,134:0) 76 %reg1040<def> = INSERT_SUBREG %reg1040, %reg1029<kill>, 13 register: %reg1040 replace range with [70,78:1) RESULT: %reg1040,0.000000e+00 = [70,78:1)[78,134:0) 0@78-(134) 1@70-(78) 84 %reg1040<def> = INSERT_SUBREG %reg1040, %reg1030<kill>, 12 register: %reg1040 replace range with [78,86:2) RESULT: %reg1040,0.000000e+00 = [70,78:1)[78,86:2)[86,134:0) 0@86-(134) 1@70-(78) 2@78-(86) 92 %reg1040<def> = INSERT_SUBREG %reg1040, %reg1031<kill>, 11 register: %reg1040 replace range with [86,94:3) RESULT: %reg1040,0.000000e+00 = [70,78:1)[78,86:2)[86,94:3)[94,134:0) 0@94-(134) 1@70-(78) 2@78-(86) 3@86-(94) rdar://problem/8096390 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@106152 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-03Use readsWritesVirtualRegister instead of counting uses and defs when insertingJakob Stoklund Olesen
spills and reloads. This means that a partial define of a register causes a reload so the other parts of the register are preserved. The reload can be prevented by adding an <imp-def> operand for the full register. This is already done by the coalescer and live interval analysis where relevant. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@105369 91177308-0d34-0410-b5e6-96231b3b80d8
2010-05-21- Change MachineInstr::findRegisterDefOperandIdx so it can also look for defsEvan Cheng
that are aliases of the specified register. - Rename modifiesRegister to definesRegister since it's looking a def of the specific register or one of its super-registers. It's not looking for def of a sub-register or alias that could change the specified register. - Added modifiesRegister to look for defs of aliases. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@104377 91177308-0d34-0410-b5e6-96231b3b80d8
2010-05-21Revert "Use MachineInstr::readsWritesVirtualRegister to determine if a ↵Jakob Stoklund Olesen
register is read." This reverts r104322. I think it was causing miscompilations. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@104323 91177308-0d34-0410-b5e6-96231b3b80d8
2010-05-21Use MachineInstr::readsWritesVirtualRegister to determine if a register is read.Jakob Stoklund Olesen
This correctly handles partial redefines and undef uses. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@104322 91177308-0d34-0410-b5e6-96231b3b80d8
2010-05-21If the first definition of a virtual register is a partial redef, add anJakob Stoklund Olesen
<imp-def> operand for the full register. This ensures that the full physical register is marked live after register allocation. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@104320 91177308-0d34-0410-b5e6-96231b3b80d8
2010-05-20Fix typo in comment.Nick Lewycky
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@104209 91177308-0d34-0410-b5e6-96231b3b80d8
2010-05-17Yes, if the redef is a copy, update the old val# with the copy. But make ↵Evan Cheng
sure to clear the copy field if the redef is not a copy. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@103922 91177308-0d34-0410-b5e6-96231b3b80d8
2010-05-15improve portability to systems that don't have powf/modf (e.g. solaris 9)Chris Lattner
patch by Evzen Muller! git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@103876 91177308-0d34-0410-b5e6-96231b3b80d8
2010-05-15A partial re-def instruction may be a copy.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@103850 91177308-0d34-0410-b5e6-96231b3b80d8
2010-05-10Re-defined valno is always valno even for partial re-def's.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@103410 91177308-0d34-0410-b5e6-96231b3b80d8
2010-05-05Teach liveintervalanalysis about virtual registers which are defined by ↵Evan Cheng
reg_sequence instructions that are formed by registers defined by distinct instructions. e.g. 80 %reg1041:6<def> = VSHRNv4i16 %reg1034<kill>, 12, pred:14, pred:%reg0 . . . 120 %reg1041:5<def> = VSHRNv4i16 %reg1039<kill>, 12, pred:14, pred:%reg0 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@103102 91177308-0d34-0410-b5e6-96231b3b80d8
2010-05-04Teach PHI elimination to remove REG_SEQUENCE instructions and update ↵Evan Cheng
references of the source operands with references of the destination with subreg indices. e.g. %reg1029<def>, %reg1030<def> = VLD1q16 %reg1024<kill>, ... %reg1031<def> = REG_SEQUENCE %reg1029<kill>, 5, %reg1030<kill>, 6 => %reg1031:5<def>, %reg1031:6<def> = VLD1q16 %reg1024<kill>, ... PHI elimination now does more than phi elimination. It is really a de-SSA pass. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@103039 91177308-0d34-0410-b5e6-96231b3b80d8
2010-05-03Print basic block numbers in live interval debug output. Since the rest of theBob Wilson
debug output is showing machine instructions, the IR-level basic block names aren't very meaningful, and because multiple machine basic blocks may be derived from one IR-level BB, they're also not unique. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@102960 91177308-0d34-0410-b5e6-96231b3b80d8
2010-04-28- Really preserve dbg_value instructions when the register is spilled.Evan Cheng
- Also, update dbg_value is the value is being re-matted from a frame slot, e.g. fixed slots for arguments. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@102565 91177308-0d34-0410-b5e6-96231b3b80d8
2010-04-26Re-enable 102323 with fix: do not update dbg_value's with incorrect frame ↵Evan Cheng
indices when the live interval are being re-materialized. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@102361 91177308-0d34-0410-b5e6-96231b3b80d8
2010-04-26Temporary disable spiller modifying dbg_value. It's breaking build.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@102327 91177308-0d34-0410-b5e6-96231b3b80d8
2010-04-26- Move TargetLowering::EmitTargetCodeForFrameDebugValue to TargetInstrInfo ↵Evan Cheng
and rename it to emitFrameIndexDebugValue. - Teach spiller to modify DBG_VALUE instructions to reference spill slots. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@102323 91177308-0d34-0410-b5e6-96231b3b80d8
2010-04-13Eliminate MachineBasicBlock::const_livein_iterator and makeDan Gohman
MachineBasicBlock::livein_iterator a const_iterator, because clients shouldn't ever be using the iterator interface to mutate the livein set. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@101147 91177308-0d34-0410-b5e6-96231b3b80d8
2010-04-07rename llvm::llvm_report_error -> llvm::report_fatal_errorChris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@100709 91177308-0d34-0410-b5e6-96231b3b80d8
2010-03-30Introduce SpecificBumpPtrAllocator, a wrapper for BumpPtrAllocator which allowsBenjamin Kramer
only a single type of object to be allocated. Use it to make VNInfo destruction typesafe. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@99919 91177308-0d34-0410-b5e6-96231b3b80d8
2010-03-30Reapply r99881 with some fixes: only call destructor in releaseMemory!Torok Edwin
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@99883 91177308-0d34-0410-b5e6-96231b3b80d8
2010-03-30Revert 99881, it brooke smooshlab's llvm-gcc-i386-darwin9.Torok Edwin
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@99882 91177308-0d34-0410-b5e6-96231b3b80d8
2010-03-30Introduce another Reset() method in BumpPtrAllocator that calls a destructorTorok Edwin
on all objects it has allocated, if they are all of the same size and alignment. Use this to destruct all VNInfos allocated in LiveIntervalAnalysis (PR6653). valnos is not reliable for this purpose, as seen in r99400 (which still leaked, and sometimes caused double frees). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@99881 91177308-0d34-0410-b5e6-96231b3b80d8
2010-03-30Avoid being influenced by the presence of dbg_value instructions.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@99879 91177308-0d34-0410-b5e6-96231b3b80d8
2010-03-24Revert Edwin's change that is breaking MultiSource/Applications/ClamAV/clamscan.Bob Wilson
--- Reverse-merging r99400 into '.': D test/CodeGen/Generic/2010-03-24-liveintervalleak.ll U lib/CodeGen/LiveIntervalAnalysis.cpp git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@99419 91177308-0d34-0410-b5e6-96231b3b80d8
2010-03-24Fix memory leak in liveintervals: the destructor for VNInfos must be called,Torok Edwin
otherwise the SmallVector it contains doesn't free its memory. In most cases LiveIntervalAnalysis could get away by not calling the destructor, because VNInfos are bumpptr-allocated, and smallvectors usually don't grow. However when the SmallVector does grow it always leaks. This is the valgrind shown leak from the original testcase: ==8206== 18,304 bytes in 151 blocks are definitely lost in loss record 164 of 164 ==8206== at 0x4A079C7: operator new(unsigned long) (vg_replace_malloc.c:220) ==8206== by 0x4DB7A7E: llvm::SmallVectorBase::grow_pod(unsigned long, unsigned long) (in /home/edwin/clam/git/builds/defaul t/libclamav/.libs/libclamav.so.6.1.0) ==8206== by 0x4F90382: llvm::VNInfo::addKill(llvm::SlotIndex) (in /home/edwin/clam/git/builds/default/libclamav/.libs/libcl amav.so.6.1.0) ==8206== by 0x5126B5C: llvm::LiveIntervals::handleVirtualRegisterDef(llvm::MachineBasicBlock*, llvm::ilist_iterator<llvm::M achineInstr>, llvm::SlotIndex, llvm::MachineOperand&, unsigned int, llvm::LiveInterval&) (in /home/edwin/clam/git/builds/defau lt/libclamav/.libs/libclamav.so.6.1.0) ==8206== by 0x512725E: llvm::LiveIntervals::handleRegisterDef(llvm::MachineBasicBlock*, llvm::ilist_iterator<llvm::MachineI nstr>, llvm::SlotIndex, llvm::MachineOperand&, unsigned int) (in /home/edwin/clam/git/builds/default/libclamav/.libs/libclamav .so.6.1.0) ==8206== by 0x51278A8: llvm::LiveIntervals::computeIntervals() (in /home/edwin/clam/git/builds/default/libclamav/.libs/libc lamav.so.6.1.0) ==8206== by 0x5127CB4: llvm::LiveIntervals::runOnMachineFunction(llvm::MachineFunction&) (in /home/edwin/clam/git/builds/de fault/libclamav/.libs/libclamav.so.6.1.0) ==8206== by 0x4DAE935: llvm::FPPassManager::runOnFunction(llvm::Function&) (in /home/edwin/clam/git/builds/default/libclama v/.libs/libclamav.so.6.1.0) ==8206== by 0x4DAEB10: llvm::FunctionPassManagerImpl::run(llvm::Function&) (in /home/edwin/clam/git/builds/default/libclama v/.libs/libclamav.so.6.1.0) ==8206== by 0x4DAED3D: llvm::FunctionPassManager::run(llvm::Function&) (in /home/edwin/clam/git/builds/default/libclamav/.l ibs/libclamav.so.6.1.0) ==8206== by 0x4D8BE8E: llvm::JIT::runJITOnFunctionUnlocked(llvm::Function*, llvm::MutexGuard const&) (in /home/edwin/clam/git/builds/default/libclamav/.libs/libclamav.so.6.1.0) ==8206== by 0x4D8CA72: llvm::JIT::getPointerToFunction(llvm::Function*) (in /home/edwin/clam/git/builds/default/libclamav/.libs/libclamav.so.6.1.0) git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@99400 91177308-0d34-0410-b5e6-96231b3b80d8
2010-03-16Fix liveintervals handling of dbg_value instructions.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@98686 91177308-0d34-0410-b5e6-96231b3b80d8
2010-03-11The check for coalescing a virtual register to a physical register, e.g.Evan Cheng
cl = EXTRACT_SUBREG reg1024, 1, is overly conservative. It should check for overlaps of vr's live interval with the super registers of the physical register (ECX in this case) and let JoinIntervals() handle checking the coalescing feasibility against the physical register (cl in this case). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@98251 91177308-0d34-0410-b5e6-96231b3b80d8
2010-03-10Fix a bug in DEBUG_VALUE handling Devang ran into.Dale Johannesen
I'll get this loop right yet. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@98155 91177308-0d34-0410-b5e6-96231b3b80d8
2010-03-01Use methods to determine if a LiveInterval is spillable.Jakob Stoklund Olesen
Don't accidentally produce unspillable intervals for deeply nested loops. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@97496 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-23Keep track of phi join registers explicitly in LiveVariables.Jakob Stoklund Olesen
Previously, LiveIntervalAnalysis would infer phi joins by looking for multiply defined registers. That doesn't work if the phi join is implicitly defined in all but one of the predecessors. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96994 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-18Always normalize spill weights, also for intervals created by spilling.Jakob Stoklund Olesen
Moderate the weight given to very small intervals. The spill weight given to new intervals created when spilling was not normalized in the same way as the original spill weights calculated by CalcSpillWeights. That meant that restored registers would tend to hang around because they had a much higher spill weight that unspilled registers. This improves the runtime of a few tests by up to 10%, and there are no significant regressions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96613 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-10When I rewrote this loop per Chris' preference IDale Johannesen
changed its behavior. Oops. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@95811 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-10Rewrite loop to suit Chris' preference.Dale Johannesen
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@95749 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-10Skip DBG_VALUE many places in live intervals andDale Johannesen
register coalescing. This fixes many crashes and places where debug info affects codegen (when dbg.value is lowered to machine instructions, which it isn't yet in TOT). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@95739 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-09move target-independent opcodes out of TargetInstrInfoChris Lattner
into TargetOpcodes.h. #include the new TargetOpcodes.h into MachineInstr. Add new inline accessors (like isPHI()) to MachineInstr, and start using them throughout the codebase. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@95687 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-06Run codegen dce pass for all targets at all optimization levels. Previously it'sEvan Cheng
only run for x86 with fastisel. I've found it being very effective in eliminating some obvious dead code as result of formal parameter lowering especially when tail call optimization eliminated the need for some of the loads from fixed frame objects. It also shrinks a number of the tests. A couple of tests no longer make sense and are now eliminated. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@95493 91177308-0d34-0410-b5e6-96231b3b80d8
2010-01-22Ignore DEBUG_VALUE when building live intervals;Dale Johannesen
this makes the code work transparently the same whether they're there or not. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@94240 91177308-0d34-0410-b5e6-96231b3b80d8
2010-01-04Change errs() to dbgs().David Greene
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@92529 91177308-0d34-0410-b5e6-96231b3b80d8
2009-12-22Changed slot index ranges for MachineBasicBlocks to be exclusive of endpoint.Lang Hames
This fixes an in-place update bug where code inserted at the end of basic blocks may not be covered by existing intervals which were live across the entire block. It is also consistent with the way ranges are specified for live intervals. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@91859 91177308-0d34-0410-b5e6-96231b3b80d8
2009-12-16Reuse lowered phi nodes.Jakob Stoklund Olesen
Tail duplication produces lots of identical phi nodes in different basic blocks. Teach PHIElimination to reuse the join registers when lowering a phi node that is identical to an already lowered node. This saves virtual registers, and more importantly it avoids creating copies the the coalescer doesn't know how to eliminate. Teach LiveIntervalAnalysis about the phi joins with multiple uses. This patch significantly reduces code size produced by -pre-regalloc-taildup. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@91549 91177308-0d34-0410-b5e6-96231b3b80d8
2009-12-11Coalesce insert_subreg undef, x first to avoid phase ordering issue.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@91103 91177308-0d34-0410-b5e6-96231b3b80d8
2009-12-10Also attempt trivial coalescing for live intervals that end in a copy.Jakob Stoklund Olesen
The coalescer is supposed to clean these up, but when setting up parameters for a function call, there may be copies to physregs. If the defining instruction has been LICM'ed far away, the coalescer won't touch it. The register allocation hint does not always work - when the register allocator is backtracking, it clears the hints. This patch is more conservative than r90502, and does not break 483.xalancbmk/i686. It still breaks the PowerPC bootstrap, so it is disabled by default, and can be enabled with the -trivial-coalesce-ends option. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@91049 91177308-0d34-0410-b5e6-96231b3b80d8
2009-12-09Added a new "splitting" spiller.Lang Hames
When a call is placed to spill an interval this spiller will first try to break the interval up into its component values. Single value intervals and intervals which have already been split (or are the result of previous splits) are spilled by the default spiller. Splitting intervals as described above may improve the performance of generated code in some circumstances. This work is experimental however, and it still miscompiles many benchmarks. It's not recommended for general use yet. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90951 91177308-0d34-0410-b5e6-96231b3b80d8
2009-12-05Temporarily revert r90502. It was causing the llvm-gcc bootstrap on PPC to fail.Bill Wendling
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90653 91177308-0d34-0410-b5e6-96231b3b80d8
2009-12-04Also attempt trivial coalescing for live intervals that end in a copy.Jakob Stoklund Olesen
The coalescer is supposed to clean these up, but when setting up parameters for a function call, there may be copies to physregs. If the defining instruction has been LICM'ed far away, the coalescer won't touch it. The register allocation hint does not always work - when the register allocator is backtracking, it clears the hints. This patch takes care of a few more cases that r90163 missed. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90502 91177308-0d34-0410-b5e6-96231b3b80d8
2009-12-03Clean up some loop logic.Jakob Stoklund Olesen
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90481 91177308-0d34-0410-b5e6-96231b3b80d8
2009-12-01Fix PR5391: support early clobber physical register def tied with a use (ewwww)Evan Cheng
- A valno should be set HasRedefByEC if there is an early clobber def in the middle of its live ranges. It should not be set if the def of the valno is defined by an early clobber. - If a physical register def is tied to an use and it's an early clobber, it just means the HasRedefByEC is set since it's still one continuous live range. - Add a couple of missing checks for HasRedefByEC in the coalescer. In general, it should not coalesce a vr with a physical register if the physical register has a early clobber def somewhere. This is overly conservative but that's the price for using such a nasty inline asm "feature". git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90269 91177308-0d34-0410-b5e6-96231b3b80d8
generated by cgit v1.2.3-18-g5258 (git 2.32.0) at 2025-05-01 20:16:07 +0000