Commit Graph

10142 Commits

Author SHA1 Message Date
Dan Gohman a8bacea11b Delete some obsolete declarations.
llvm-svn: 83856
2009-10-12 16:43:44 +00:00
Chris Lattner 40cf28d6eb add some more hooks to the C bindings, patch by Kenneth Uildriks!
llvm-svn: 83821
2009-10-12 04:01:02 +00:00
Edward O'Callaghan 396ed2bcaa Fix LLVM CMake build system so that it may now work on Solaris and AuroraUX.
llvm-svn: 83819
2009-10-12 04:00:11 +00:00
Nick Lewycky c7cf0b223b Fix typo.
llvm-svn: 83780
2009-10-11 18:53:09 +00:00
Nick Lewycky 5246987a76 Fix typo.
llvm-svn: 83779
2009-10-11 18:47:33 +00:00
Chris Lattner d1d23e1949 add a helper for matching "1".
llvm-svn: 83760
2009-10-11 07:51:25 +00:00
Chris Lattner 09677258e5 there is no need to run mem2reg after jump threading at LTO time now.
llvm-svn: 83753
2009-10-11 04:17:33 +00:00
Chris Lattner bb094e12f2 fix a bunch of bad formatting, delete the dead
ConstantInt::TheTrueVal/TheFalseVal members.

llvm-svn: 83752
2009-10-11 04:03:22 +00:00
Chris Lattner 9c382cebc5 add a simple helper method.
llvm-svn: 83745
2009-10-10 23:41:48 +00:00
Chris Lattner 249265de06 add ability for clients of SSAUpdater to find out about the
PHI nodes inserted.

llvm-svn: 83744
2009-10-10 23:15:24 +00:00
Chris Lattner 67cdd8b567 add the ability to get a rewritten value from the middle of a block,
not just at the end.  Add a big comment explaining when this could
be useful (which never happens for jump threading).

llvm-svn: 83741
2009-10-10 23:00:11 +00:00
Chris Lattner e474a8d3a7 rename GetValueInBlock -> GetValueAtEndOfBlock to better reflect
what it does.

llvm-svn: 83740
2009-10-10 22:41:58 +00:00
Chris Lattner 60d4e69c81 Implement an efficient and fully general SSA update mechanism that
works on unstructured CFGs.  This implements PR217, our oldest open PR.

llvm-svn: 83705
2009-10-10 09:04:27 +00:00
Chris Lattner 71db463fcf add some WeakVH::operator='s. Without these, assigning
a Value* to a WeakVH was constructing a temporary WeakVH
(due to the implicit assignment operator).  This avoids
that cost.

llvm-svn: 83704
2009-10-10 08:27:29 +00:00
Chris Lattner a0424f5372 change some static_cast into cast, pointed out by Gabor.
llvm-svn: 83703
2009-10-10 08:01:27 +00:00
Chris Lattner 99e02edef8 add a version of PHINode::getIncomingBlock that takes a raw
Use, to complement the version that takes a use_iterator.

llvm-svn: 83702
2009-10-10 07:42:42 +00:00
Dan Gohman a22f2d8614 Make getMachineNode return a MachineSDNode* instead of a generic SDNode*
since it won't do any folding. This will help avoid some inconvenient
casting.

llvm-svn: 83698
2009-10-10 01:29:16 +00:00
Dan Gohman e919de5acf Replace X86's CanRematLoadWithDispOperand by calling the target-independent
MachineInstr::isInvariantLoad instead, which has the benefit of being
more complete.

llvm-svn: 83696
2009-10-10 00:34:18 +00:00
Dan Gohman 918ec53c64 The ScheduleDAG framework now requires an AliasAnalysis argument, though
it isn't needed in the ScheduleDAGSDNodes schedulers.

llvm-svn: 83691
2009-10-09 23:33:48 +00:00
Dan Gohman 87b02d5bbc Factor out LiveIntervalAnalysis' code to determine whether an instruction
is trivially rematerializable and integrate it into
TargetInstrInfo::isTriviallyReMaterializable. This way, all places that
need to know whether an instruction is rematerializable will get the
same answer.

This enables the useful parts of the aggressive-remat option by
default -- using AliasAnalysis to determine whether a memory location
is invariant, and removes the questionable parts -- rematting operations
with virtual register inputs that may not be live everywhere.

llvm-svn: 83687
2009-10-09 23:27:56 +00:00
Devang Patel df45c7f642 Extract scope information from the variable itself, instead of relying on alloca or llvm.dbg.declare location.
While recording beginning of a function, use scope info from the first location entry instead of just relying on first location entry itself.

llvm-svn: 83684
2009-10-09 22:42:28 +00:00
Jeffrey Yasskin 307c053f2e ExecutionEngine::clearGlobalMappingsFromModule failed to remove reverse
mappings, which could cause errors and assert-failures.  This patch fixes that,
adds a test, and refactors the global-mapping-removal code into a single place.

llvm-svn: 83678
2009-10-09 22:10:27 +00:00
Dan Gohman f4eb6777e5 Add a const qualifier.
llvm-svn: 83677
2009-10-09 22:09:05 +00:00
Dale Johannesen 96a5b87ae2 Use names instead of numbers for some of the magic
constants used in inlining heuristics (especially
those used in more than one file).  No functional change.

llvm-svn: 83675
2009-10-09 21:42:02 +00:00
Dan Gohman dd76bb23d1 Add basic infrastructure and x86 support for preserving MachineMemOperand
information when unfolding memory references.

llvm-svn: 83656
2009-10-09 18:10:05 +00:00
Mikhail Glushenkov fc8809d0ac Raise the limit on built-in plugins in llvmc to 10.
llvm-svn: 83614
2009-10-09 04:15:52 +00:00
Evan Cheng f5970a019e Indentation.
llvm-svn: 83607
2009-10-09 00:57:38 +00:00
Dan Gohman 816fe0a4aa Add the ability to track HasNSW and HasNUW on more kinds of SCEV expressions.
llvm-svn: 83601
2009-10-09 00:10:36 +00:00
Bob Wilson 2a45a65511 Add a SelectionDAG getTargetInsertSubreg convenience function,
similar to getTargetExtractSubreg.

llvm-svn: 83564
2009-10-08 18:49:46 +00:00
Dan Gohman 09984279fd Add a form of addPreserved which takes a string argument, to allow passes
to declare that they preserve other passes without needing to pull in
additional header file or library dependencies. Convert MachineFunctionPass
and CodeGenLICM to make use of this.

llvm-svn: 83555
2009-10-08 17:00:02 +00:00
Chris Lattner 78765deded remove LoopVR pass. According to Nick:
"LoopVR's logic was copied into ScalarEvolution::getUnsignedRange and 
::getSignedRange. Please delete LoopVR."

llvm-svn: 83531
2009-10-08 06:42:44 +00:00
Jim Grosbach c0615aa17f Re-enable register scavenging in Thumb1 by default.
llvm-svn: 83521
2009-10-08 01:46:59 +00:00
Jim Grosbach 456735c54b reverting thumb1 scavenging default due to test failure while I figure out what's up.
llvm-svn: 83501
2009-10-07 22:49:41 +00:00
Chris Lattner 63fe7ff7d9 second half of lazy liveness removal.
llvm-svn: 83500
2009-10-07 22:49:30 +00:00
Jim Grosbach 267fa622fa Enable thumb1 register scavenging by default.
llvm-svn: 83496
2009-10-07 22:26:31 +00:00
Devang Patel 55571bd7bd Extract subprogram and compile unit information from the debug info attached to an instruction.
llvm-svn: 83491
2009-10-07 22:04:08 +00:00
Dan Gohman be8137b0b4 Replace TargetInstrInfo::isInvariantLoad and its target-specific
implementations with a new MachineInstr::isInvariantLoad, which uses
MachineMemOperands and is target-independent. This brings MachineLICM
and other functionality to targets which previously lacked an
isInvariantLoad implementation.

llvm-svn: 83475
2009-10-07 17:38:06 +00:00
Jim Grosbach fa14dd430c Add register-reuse to frame-index register scavenging. When a target uses
a virtual register to eliminate a frame index, it can return that register
and the constant stored there to PEI to track. When scavenging to allocate
for those registers, PEI then tracks the last-used register and value, and
if it is still available and matches the value for the next index, reuses
the existing value rather and removes the re-materialization instructions.
Fancier tracking and adjustment of scavenger allocations to keep more
values live for longer is possible, but not yet implemented and would likely
be better done via a different, less special-purpose, approach to the
problem.

eliminateFrameIndex() is modified so the target implementations can return
the registers they wish to be tracked for reuse.

ARM Thumb1 implements and utilizes the new mechanism. All other targets are
simply modified to adjust for the changed eliminateFrameIndex() prototype.

llvm-svn: 83467
2009-10-07 17:12:56 +00:00
Torok Edwin 822f634d29 Add a comment explaining how DenseMap::insert works, because it is not
intuitive.
It does NOT update the value if the key is already in the map,
it also returns false if the key is already in the map, regardless
if the value matched.

llvm-svn: 83458
2009-10-07 09:23:56 +00:00
Duncan Sands 986742a14f Make getPointerTo return a const PointerType* rather than
an unqualified PointerType* because it seems more correct.

llvm-svn: 83454
2009-10-07 07:35:19 +00:00
Dan Gohman 2a685bdbee INTRINSIC_W_CHAIN and INTRINSIC_VOID do not use MemSDNode. They
may access memory, but they don't carry a MachineMemOperand.

llvm-svn: 83449
2009-10-07 03:00:18 +00:00
Devang Patel 4598eb6214 Add support to handle debug info attached to an instruction.
This is not yet enabled.

llvm-svn: 83400
2009-10-06 18:37:31 +00:00
Dan Gohman e66abdc58f Make LLVMContext's pImpl member const.
llvm-svn: 83393
2009-10-06 17:43:57 +00:00
Chris Lattner a893f5bdf5 remove predicate simplifier, it never got the last bugs beaten
out of it, and jump threading, condprop and gvn are now getting
most of the benefit.  This was approved by Nicholas and Nicolas.

llvm-svn: 83390
2009-10-06 16:59:46 +00:00
Duncan Sands 9ed7b16bf3 Introduce and use convenience methods for getting pointer types
where the element is of a basic builtin type.  For example, to get
an i8* use getInt8PtrTy.

llvm-svn: 83379
2009-10-06 15:40:36 +00:00
Devang Patel 051454a16f Update processDebugLoc() so that it can be used to process debug info before and after printing an instruction.
llvm-svn: 83363
2009-10-06 02:19:11 +00:00
Jim Grosbach 2dfb5da6bb In Thumb1, the register scavenger is not always able to use an emergency
spill slot. When frame references are via the frame pointer, they will be
negative, but Thumb1 load/store instructions only allow positive immediate
offsets. Instead, Thumb1 will spill to R12.

llvm-svn: 83336
2009-10-05 22:30:23 +00:00
Dan Gohman e525d9ddc0 Remove an unnnecessary LLVMContext argument in
ConstantFoldLoadThroughGEPConstantExpr.

llvm-svn: 83311
2009-10-05 16:36:26 +00:00
Dan Gohman 5d36b9c447 Fix a name in a comment.
llvm-svn: 83306
2009-10-05 15:23:17 +00:00
Chris Lattner 3a2e503e33 add more type predicates.
llvm-svn: 83296
2009-10-05 05:48:40 +00:00