aboutsummaryrefslogtreecommitdiff
path: root/lib/CodeGen/SelectionDAG
AgeCommit message (Collapse)Author
2008-07-04Rather than having a different custom legalizationDuncan Sands
hook for each way in which a result type can be legalized (promotion, expansion, softening etc), just use one: ReplaceNodeResults, which returns a node with exactly the same result types as the node passed to it, but presumably with a bunch of custom code behind the scenes. No change if the new LegalizeTypes infrastructure is not turned on. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@53137 91177308-0d34-0410-b5e6-96231b3b80d8
2008-07-03Revert my previous check-in that split up MachineModuleInfo. It turns out toBill Wendling
slow the compiler down at -O0 some 30% or more. Ooops. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@53120 91177308-0d34-0410-b5e6-96231b3b80d8
2008-07-03Backed out 53031.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@53110 91177308-0d34-0410-b5e6-96231b3b80d8
2008-07-03Avoid unnecessarily copying APInt objects.Dan Gohman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@53065 91177308-0d34-0410-b5e6-96231b3b80d8
2008-07-02Replace a few uses of SelectionDAG::getTargetNode withDan Gohman
SelectionDAG::SelectNodeTo in the instruction selector. This updates existing nodes in place instead of creating new ones. Go back to selecting ISD::DBG_LABEL nodes into TargetInstrInfo::DBG_LABEL nodes instead of leaving them unselected, now that SelectNodeTo allows us to update them in place. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@53057 91177308-0d34-0410-b5e6-96231b3b80d8
2008-07-02Add a new getMergeValues method that does not needDuncan Sands
to be passed the list of value types, and use this where appropriate. Inappropriate places are where the value type list is already known and may be long, in which case the existing method is more efficient. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@53035 91177308-0d34-0410-b5e6-96231b3b80d8
2008-07-02Fixed problem in EmitStackConvert where the source and target typeMon P Wang
have different alignment by creating a stack slot with the max alignment of source and target type. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@53031 91177308-0d34-0410-b5e6-96231b3b80d8
2008-07-02instead of aborting on shifts of i1, just implicitly fold them.Chris Lattner
The dag combiner can produce a shift of i1 when folding icmp i1's. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@53030 91177308-0d34-0410-b5e6-96231b3b80d8
2008-07-02Fix typo compounded by a cut-and-pasto.Duncan Sands
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@53012 91177308-0d34-0410-b5e6-96231b3b80d8
2008-07-02Let AnalyzeNewNode take care of calling ExpungeNode.Duncan Sands
This makes sure that all new nodes are expunged, not just those the top node of a new subtree. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@53011 91177308-0d34-0410-b5e6-96231b3b80d8
2008-07-02- Use a faster priority comparison function if -fast.Evan Cheng
- Code clean up. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@53010 91177308-0d34-0410-b5e6-96231b3b80d8
2008-07-01No need to use std::distance. We can just count the number of operandsOwen Anderson
much more cheaply. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52990 91177308-0d34-0410-b5e6-96231b3b80d8
2008-07-01Eliminate a compile time warning.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52982 91177308-0d34-0410-b5e6-96231b3b80d8
2008-07-01Do run ComputeLiveOutVRegInfo with -fast.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52975 91177308-0d34-0410-b5e6-96231b3b80d8
2008-07-01Do not use computationally expensive scheduling heuristics with -fast.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52971 91177308-0d34-0410-b5e6-96231b3b80d8
2008-07-01Apply Chris' suggestion.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52970 91177308-0d34-0410-b5e6-96231b3b80d8
2008-07-01Split ISD::LABEL into ISD::DBG_LABEL and ISD::EH_LABEL, eliminatingDan Gohman
the need for a flavor operand, and add a new SDNode subclass, LabelSDNode, for use with them to eliminate the need for a label id operand. Change instruction selection to let these label nodes through unmodified instead of creating copies of them. Teach the MachineInstr emitter how to emit a MachineInstr directly from an ISD label node. This avoids the need for allocating SDNodes for the label id and flavor value, as well as SDNodes for each of the post-isel label, label id, and label flavor. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52943 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-30Suppress compiler warning.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52934 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-30Use a simpler but equivalent form of RecordSource.Dan Gohman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52931 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-30Add timing report for various sub-passes under SelectionDAGISel.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52930 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-30Use reserve.Dan Gohman
SelectionDAG::allnodes_size is linear, but that doesn't appear to outweigh the benefit of reducing heap traffic. If it does become a problem, we should teach SelectionDAG to keep a count of how many nodes are live, because there are several other places where that information would be useful as well. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52926 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-30Rename ISD::LOCATION to ISD::DBG_STOPPOINT to better reflect itsDan Gohman
purpose, and give it a custom SDNode subclass so that it doesn't need to have line number, column number, filename string, and directory string, all existing as individual SDNodes to be the operands. This was the only user of ISD::STRING, StringSDNode, etc., so remove those and some associated code. This makes stop-points considerably easier to read in -view-legalize-dags output, and reduces overhead (creating new nodes and copying std::strings into them) on code containing debugging information. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52924 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-30Split scheduling from instruction selection.Evan Cheng
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52923 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-30Replace some std::vectors that showed up in heap profiling withDan Gohman
SmallVectors. Change the signature of TargetLowering::LowerArguments to avoid returning a vector by value, and update the two targets which still use this directly, Sparc and IA64, accordingly. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52917 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-30Correct the allocation size for CCState's UsedRegs member, whichDan Gohman
only needs one bit for each register. UsedRegs is a SmallVector sized at 16, so this eliminates a heap allocation/free for every call and return processed by Legalize on most targets. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52915 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-30ExpungeNode is only needed for new nodes! ThisDuncan Sands
fixes CodeGen/PowerPC/2008-06-19-LegalizerCrash.ll when using the new LegalizeTypes infrastructure. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52903 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-30Support for VAARG. As noted in a comment, this isDuncan Sands
wrong for types like x86 long double and i1, but no worse than what is done in LegalizeDAG. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52898 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-30Support for promoting select_cc operands.Duncan Sands
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52895 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-30Revert the SelectionDAG optimization that makesDuncan Sands
it impossible to create a MERGE_VALUES node with only one result: sometimes it is useful to be able to create a node with only one result out of one of the results of a node with more than one result, for example because the new node will eventually be used to replace a one-result node using ReplaceAllUsesWith, cf X86TargetLowering::ExpandFP_TO_SINT. On the other hand, most users of MERGE_VALUES don't need this and for them the optimization was valuable. So add a new utility method getMergeValues for creating MERGE_VALUES nodes which by default performs the optimization. Change almost everywhere to use getMergeValues (and tidy some stuff up at the same time). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52893 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-30- Re-apply 52748 and friends with fix. GetConstantStringInfo() returns an ↵Evan Cheng
empty string for ConstantAggregateZero case which surprises selectiondag. - Correctly handle memcpy from constant string which is zero-initialized. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52891 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-30Implement split and scalarize for SELECT_CC, fixing PR2504Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52887 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-29Revert (52748 and friends):Anton Korobeynikov
Move GetConstantStringInfo to lib/Analysis. Remove string output routine from Constant. Update all callers. Change debug intrinsic api slightly to accomodate move of routine, these now return values instead of strings. This unbreaks llvm-gcc bootstrap. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52884 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-28Really fix the bootstrap failure.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52854 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-28Add back the capability to include nul characters in strings with Chris Lattner
GetConstantStringInfo. This will hopefully restore llvm-gcc to happy bootstrap land. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52851 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-28When folding a bitcast into a load or store, preserve the alignmentDan Gohman
information of the original load or store, which is checked to be at least as good, and possibly better. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52849 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-27simplify this check, GetConstantStringInfo validates that aChris Lattner
global is constant already. No functionality change. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52812 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-27Refactor the DebugInfoDesc stuff out of the MachineModuleInfo file. Clean upBill Wendling
some uses of std::vector, where it's return std::vector by value. Yuck! git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52800 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-26duncan points out that isOperationLegal includes a check forChris Lattner
type legality. Thanks Duncan! git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52786 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-26Move GetConstantStringInfo to lib/Analysis. RemoveEric Christopher
string output routine from Constant. Update all callers. Change debug intrinsic api slightly to accomodate move of routine, these now return values instead of strings. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52748 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-26when we know the signbit of an input to uint_to_fp is zero,Chris Lattner
change it to sint_to_fp on targets where that is cheaper (and visaversa of course). This allows us to compile uint_to_fp to: _test: movl 4(%esp), %eax shrl $23, %eax cvtsi2ss %eax, %xmm0 movl 8(%esp), %eax movss %xmm0, (%eax) ret instead of: .align 3 LCPI1_0: ## double .long 0 ## double least significant word 4.5036e+15 .long 1127219200 ## double most significant word 4.5036e+15 .text .align 4,0x90 .globl _test _test: subl $12, %esp movl 16(%esp), %eax shrl $23, %eax movl %eax, (%esp) movl $1127219200, 4(%esp) movsd (%esp), %xmm0 subsd LCPI1_0, %xmm0 cvtsd2ss %xmm0, %xmm0 movl 20(%esp), %eax movss %xmm0, (%eax) addl $12, %esp ret git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52747 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-25- Fix a x86 vector isel bug: illegal transformation of a vector_shuffle into aEvan Cheng
shift. - Add a readme entry for a missing vector_shuffle optimization that results in awful codegen. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52740 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-25Add support for expanding PPC 128 bit floats.Duncan Sands
For this it is convenient to permit floats to be used with EXTRACT_ELEMENT, so I tweaked things to allow that. I also added libcalls for ppcf128 to i32 forms of FP_TO_XINT, since they exist in libgcc and this case can certainly occur (and does occur in the testsuite) - before the i64 libcall was being used. Also, the XINT_TO_FP result seemed to be wrong when the argument is an i128: the wrong fudge factor was added (the i32 and i64 cases were handled directly, but the i128 code fell through to some generic softening code which seemed to think it was i64 to f32!). So I fixed it by adding a fudge factor that I found in my breakfast cereal. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52739 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-25Add/complete support for integer and floatDuncan Sands
select_cc and friends. This code could be factorized a bit but I'm not sure that it's worth it. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52724 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-25Remove the OrigVT member from AtomicSDNode, as it is redundant withDan Gohman
the base SDNode's VTList. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52722 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-25Added MemOperands to Atomic operations since Atomics touches memory.Mon P Wang
Added abstract class MemSDNode for any Node that have an associated MemOperand Changed atomic.lcs => atomic.cmp.swap, atomic.las => atomic.load.add, and atomic.lss => atomic.load.sub git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52706 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-23Use the new PriorityQueue in ScheduleDAGList too, which alsoDan Gohman
needs arbitrary-element removal. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52654 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-23Move a DenseMap's declaration outside of a loop, and just callDan Gohman
clear() on each iteration. This avoids allocating and deallocating all of DenseMap's memory on each iteration. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52642 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-23Duncan pointed out this code could be tidied.Dan Gohman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52624 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-23Port some integer multiplication fixes from LegalizeDAG.Duncan Sands
Bail out with an error if there is no libcall available for the given size of integer. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52622 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-23Support for expanding the result of EXTRACT_ELEMENT.Duncan Sands
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52621 91177308-0d34-0410-b5e6-96231b3b80d8
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2026-03-22 18:22:58 +0000