llvm-6502/lib/Analysis
Hal Finkel db1e51359d Revert r219432 - "Revert "[BasicAA] Revert "Revert r218714 - Make better use of zext and sign information."""
Let's try this again...

This reverts r219432, plus a bug fix.

Description of the bug in r219432 (by Nick):

The bug was using AllPositive to break out of the loop; if the loop break
condition i != e is changed to i != e && AllPositive then the
test_modulo_analysis_with_global test I've added will fail as the Modulo will
be calculated incorrectly (as the last loop iteration is skipped, so Modulo
isn't updated with its Scale).

Nick also adds this comment:

ComputeSignBit is safe to use in loops as it takes into account phi nodes, and
the  == EK_ZeroEx check is safe in loops as, no matter how the variable changes
between iterations, zero-extensions will always guarantee a zero sign bit. The
isValueEqualInPotentialCycles check is therefore definitely not needed as all
the variable analysis holds no matter how the variables change between loop
iterations.

And this patch also adds another enhancement to GetLinearExpression - basically
to convert ConstantInts to Offsets (see test_const_eval and
test_const_eval_scaled for the situations this improves).

Original commit message:

This reverts r218944, which reverted r218714, plus a bug fix.

Description of the bug in r218714 (by Nick):

The original patch forgot to check if the Scale in VariableGEPIndex flipped the
sign of the variable. The BasicAA pass iterates over the instructions in the
order they appear in the function, and so BasicAliasAnalysis::aliasGEP is
called with the variable it first comes across as parameter GEP1. Adding a
%reorder label puts the definition of %a after %b so aliasGEP is called with %b
as the first parameter and %a as the second. aliasGEP later calculates that %a
== %b + 1 - %idxprom where %idxprom >= 0 (if %a was passed as the first
parameter it would calculate %b == %a - 1 + %idxprom where %idxprom >= 0) -
ignoring that %idxprom is scaled by -1 here lead the patch to incorrectly
conclude that %a > %b.

Revised patch by Nick White, thanks! Thanks to Lang to isolating the bug.
Slightly modified by me to add an early exit from the loop and avoid
unnecessary, but expensive, function calls.

Original commit message:

Two related things:

 1. Fixes a bug when calculating the offset in GetLinearExpression. The code
    previously used zext to extend the offset, so negative offsets were converted
    to large positive ones.

 2. Enhance aliasGEP to deduce that, if the difference between two GEP
    allocations is positive and all the variables that govern the offset are also
    positive (i.e. the offset is strictly after the higher base pointer), then
    locations that fit in the gap between the two base pointers are NoAlias.

Patch by Nick White!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221876 91177308-0d34-0410-b5e6-96231b3b80d8
2014-11-13 09:16:54 +00:00
..
IPA CGSCC should not treat intrinsic calls like function calls (PR21403) 2014-11-12 18:25:47 +00:00
AliasAnalysis.cpp Reformat partially, where I touched for whitespace changes. 2014-10-28 11:54:52 +00:00
AliasAnalysisCounter.cpp
AliasAnalysisEvaluator.cpp
AliasDebugger.cpp
AliasSetTracker.cpp Use AA in LoadCombine 2014-11-03 23:19:16 +00:00
Analysis.cpp
AssumptionTracker.cpp Clean up assume intrinsic pattern matching, no need to check that the argument is a value. 2014-10-25 18:09:01 +00:00
BasicAliasAnalysis.cpp Revert r219432 - "Revert "[BasicAA] Revert "Revert r218714 - Make better use of zext and sign information.""" 2014-11-13 09:16:54 +00:00
BlockFrequencyInfo.cpp
BlockFrequencyInfoImpl.cpp [modules] Stop excluding Support/Debug.h from the Support module. This header 2014-10-13 00:41:03 +00:00
BranchProbabilityInfo.cpp Revert "IR: MDNode => Value" 2014-11-11 21:30:22 +00:00
CaptureTracking.cpp
CFG.cpp
CFGPrinter.cpp
CFLAliasAnalysis.cpp [CFL-AA] CFL-AA should not assert on an va_arg instruction 2014-10-14 20:51:26 +00:00
CGSCCPassManager.cpp
CMakeLists.txt Add file to CMake build as well. 2014-09-18 00:39:20 +00:00
CodeMetrics.cpp Treat the WorkSet used to find ephemeral values as double-ended 2014-10-15 17:34:48 +00:00
ConstantFolding.cpp [InstSimplify] Support constant folding to vector of pointers 2014-10-22 12:18:48 +00:00
CostModel.cpp
Delinearization.cpp
DependenceAnalysis.cpp Reformat partially, where I touched for whitespace changes. 2014-10-28 11:54:52 +00:00
DominanceFrontier.cpp
DomPrinter.cpp
FunctionTargetTransformInfo.cpp Add a new pass FunctionTargetTransformInfo. This pass serves as a 2014-09-18 00:34:14 +00:00
InstCount.cpp
InstructionSimplify.cpp InstSimplify: Exact shifts of X by Y are X if X has the lsb set 2014-11-05 00:59:59 +00:00
Interval.cpp
IntervalPartition.cpp
IVUsers.cpp
JumpInstrTableInfo.cpp Add Forward Control-Flow Integrity. 2014-11-11 21:08:02 +00:00
LazyCallGraph.cpp
LazyValueInfo.cpp [LVI] Add some additional comments about caching and context instructions 2014-10-16 00:40:05 +00:00
LibCallAliasAnalysis.cpp
LibCallSemantics.cpp remove function names from comments; NFC 2014-10-21 18:26:57 +00:00
Lint.cpp Make use of @llvm.assume in ValueTracking (computeKnownBits, etc.) 2014-09-07 18:57:58 +00:00
LLVMBuild.txt
Loads.cpp Revert "Teach the load analysis to allow finding available values which require" (r220277) 2014-10-21 23:49:52 +00:00
LoopInfo.cpp Revert "IR: MDNode => Value" 2014-11-11 21:30:22 +00:00
LoopPass.cpp Fix assertion in LICM doFinalization() 2014-09-24 16:48:31 +00:00
Makefile
MemDepPrinter.cpp
MemoryBuiltins.cpp PR21145: Teach LLVM about C++14 sized deallocation functions. 2014-10-03 20:17:06 +00:00
MemoryDependenceAnalysis.cpp Silence a -Wsign-compare warning. NFC. 2014-10-02 13:17:11 +00:00
ModuleDebugInfoPrinter.cpp
NoAliasAnalysis.cpp Simplify code. No functionality change. 2014-10-05 12:21:57 +00:00
PHITransAddr.cpp Make use of @llvm.assume in ValueTracking (computeKnownBits, etc.) 2014-09-07 18:57:58 +00:00
PostDominators.cpp
PtrUseVisitor.cpp
README.txt
RegionInfo.cpp
RegionPass.cpp
RegionPrinter.cpp
ScalarEvolution.cpp Teach ScalarEvolution to sharpen range information. 2014-11-13 00:00:58 +00:00
ScalarEvolutionAliasAnalysis.cpp
ScalarEvolutionExpander.cpp Make use of @llvm.assume in ValueTracking (computeKnownBits, etc.) 2014-09-07 18:57:58 +00:00
ScalarEvolutionNormalization.cpp
ScopedNoAliasAA.cpp Revert "IR: MDNode => Value" 2014-11-11 21:30:22 +00:00
SparsePropagation.cpp
StratifiedSets.h
TargetTransformInfo.cpp Add minnum / maxnum intrinsics 2014-10-21 23:00:20 +00:00
Trace.cpp
TypeBasedAliasAnalysis.cpp Revert "IR: MDNode => Value" 2014-11-11 21:30:22 +00:00
ValueTracking.cpp Revert "IR: MDNode => Value" 2014-11-11 21:30:22 +00:00

Analysis Opportunities:

//===---------------------------------------------------------------------===//

In test/Transforms/LoopStrengthReduce/quadradic-exit-value.ll, the
ScalarEvolution expression for %r is this:

  {1,+,3,+,2}<loop>

Outside the loop, this could be evaluated simply as (%n * %n), however
ScalarEvolution currently evaluates it as

  (-2 + (2 * (trunc i65 (((zext i64 (-2 + %n) to i65) * (zext i64 (-1 + %n) to i65)) /u 2) to i64)) + (3 * %n))

In addition to being much more complicated, it involves i65 arithmetic,
which is very inefficient when expanded into code.

//===---------------------------------------------------------------------===//

In formatValue in test/CodeGen/X86/lsr-delayed-fold.ll,

ScalarEvolution is forming this expression:

((trunc i64 (-1 * %arg5) to i32) + (trunc i64 %arg5 to i32) + (-1 * (trunc i64 undef to i32)))

This could be folded to

(-1 * (trunc i64 undef to i32))

//===---------------------------------------------------------------------===//