1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-24 19:52:54 +01:00
Commit Graph

12092 Commits

Author SHA1 Message Date
Jakob Stoklund Olesen
71ad56676d Don't track kills in VNInfo. Use interval ends instead.
The VNInfo.kills vector was almost unused except for all the code keeping it
updated. The few places using it were easily rewritten to check for interval
ends instead.

The two new methods LiveInterval::killedAt and killedInRange are replacements.

This brings us down to 3 independent data structures tracking kills.

llvm-svn: 106905
2010-06-25 22:53:05 +00:00
Evan Cheng
346aecdb8b Change if-conversion block size limit checks to add some flexibility.
llvm-svn: 106901
2010-06-25 22:42:03 +00:00
Dale Johannesen
b1fc776fca The hasMemory argument is irrelevant to how the argument
for an "i" constraint should get lowered; PR 6309.  While
this argument was passed around a lot, this is the only
place it was used, so it goes away from a lot of other
places.

llvm-svn: 106893
2010-06-25 21:55:36 +00:00
Jim Grosbach
3aa1632045 80-column and trailing whitespace cleanup
llvm-svn: 106885
2010-06-25 21:02:24 +00:00
Benjamin Kramer
afe7228ede Tweak MemoryBuffer to allocate the class itself, the name and possibly the
buffer in the same chunk of memory.

2 less mallocs for every uninitialized MemoryBuffer and 1 less malloc for every
MemoryBuffer pointing to a memory range translate into 20% less mallocs on
clang -cc1 -Eonly Cocoa_h.m.

llvm-svn: 106839
2010-06-25 11:50:40 +00:00
Oscar Fuentes
4c30188483 DataTypes.h.cmake (MSVC): Protect some #define's and use specific
suffixes for integer types.

llvm-svn: 106797
2010-06-24 22:05:52 +00:00
Jakob Stoklund Olesen
7a5bf34236 Remove the now unused LiveIntervals::getVNInfoSourceReg().
This method was always a bit too simplistic for the real world. It didn't really
deal with subregisters and such.

llvm-svn: 106781
2010-06-24 20:18:15 +00:00
Eric Christopher
7405c7bdb8 Silence some unused variable warnings.
llvm-svn: 106778
2010-06-24 20:14:34 +00:00
Jakob Stoklund Olesen
1a6a8cfc51 Remove the -fast-spill option.
This code path has never really been used, and we are going to be handling
spilling through the Spiller interface in the future.

llvm-svn: 106777
2010-06-24 19:56:08 +00:00
Jakob Stoklund Olesen
9f8104463f Replace a big gob of old coalescer logic with the new CoalescerPair class.
CoalescerPair can determine if a copy can be coalesced, and which register gets
merged away. The old logic in SimpleRegisterCoalescing had evolved into
something a bit too convoluted.

This second attempt fixes some crashes that only occurred Linux.

llvm-svn: 106769
2010-06-24 18:15:01 +00:00
Bob Wilson
2b2d62ad11 Edit and clarify comments for TargetInstrInfo methods:
None of the existing implementations of commuteInstruction create new
instructions unless the NewMI parameter is true, but the comment had
implied otherwise.

findCommutedOpIndices returns false, not true, when it doesn't know
how to commute the instruction.

llvm-svn: 106761
2010-06-24 16:48:26 +00:00
Dan Gohman
4f088f99fe Add a comment about a thread safety issue in Statistic.h.
llvm-svn: 106756
2010-06-24 16:31:32 +00:00
Dan Gohman
abeebe353c Add overloads for getFile and getFileOrSTDIN which take a const char *
instead of a StringRef, avoiding the need to copy the string in the
common case.

llvm-svn: 106754
2010-06-24 16:25:50 +00:00
Jakob Stoklund Olesen
f01eb3aab6 Be more strict about subreg-to-subreg copies in CoalescerPair.
Also keep track of the original DstREg before subregister adjustments.

llvm-svn: 106753
2010-06-24 16:19:28 +00:00
Dan Gohman
a08a9b8a0b Reapply r106634, now that the bug it exposed is fixed.
llvm-svn: 106746
2010-06-24 14:30:44 +00:00
Jakob Stoklund Olesen
1c9d50ed92 Revert "Replace a big gob of old coalescer logic with the new CoalescerPair class."
Whiny buildbots.

llvm-svn: 106710
2010-06-24 00:52:22 +00:00
Devang Patel
0c927a80fe Use ValueMap instead of DenseMap.
The ValueMapper used by various cloning utility maps MDNodes also.

llvm-svn: 106706
2010-06-24 00:33:28 +00:00
Jakob Stoklund Olesen
19abbf4387 Replace a big gob of old coalescer logic with the new CoalescerPair class.
CoalescerPair can determine if a copy can be coalesced, and which register gets
merged away. The old logic in SimpleRegisterCoalescing had evolved into
something a bit too convoluted.

llvm-svn: 106701
2010-06-24 00:12:39 +00:00
Devang Patel
373b10cbaa Cosmetic change.
Do not use "ValueMap" as a name for a local variable or an argument.

llvm-svn: 106698
2010-06-23 23:55:51 +00:00
Gabor Greif
4615fc1c93 minor enhancement to llvm::isFreeCall API: return CallInst; no functional change
llvm-svn: 106686
2010-06-23 21:51:12 +00:00
Daniel Dunbar
be50ef88bd Revert r106263, "Fold the ShrinkDemandedOps pass into the regular DAGCombiner pass,"... it was causing both 'file' (with clang) and 176.gcc (with llvm-gcc) to be miscompiled.
llvm-svn: 106634
2010-06-23 17:09:26 +00:00
Jim Grosbach
12afda541e Some targets don't require the fencing MEMBARRIER instructions surrounding
atomic intrinsics, either because the use locking instructions for the
atomics, or because they perform the locking directly. Add support in the
DAG combiner to fold away the fences.

llvm-svn: 106630
2010-06-23 16:07:42 +00:00
Jim Grosbach
395b0a671c remove trailing whitespace
llvm-svn: 106628
2010-06-23 15:49:18 +00:00
Dan Gohman
f44783ab4e Replace ScalarEvolution's private copy of getLoopPredecessor
with LoopInfo's public copy.

llvm-svn: 106603
2010-06-22 23:43:28 +00:00
Eric Christopher
835fc55fd0 Look for and use a different darwin crash reporter library.
llvm-svn: 106576
2010-06-22 21:01:04 +00:00
Dan Gohman
2303159bba Move PHIElimination's SplitCriticalEdge for MachineBasicBlocks out
into a utility routine, teach it how to update MachineLoopInfo, and
make use of it in MachineLICM to split critical edges on demand.

llvm-svn: 106555
2010-06-22 17:25:57 +00:00
Dan Gohman
29b845f3b7 Split out the code for finding a unique loop predecessor from
getLoopPreheader into a separate function, for clients which don't
require a proper preheader.

llvm-svn: 106543
2010-06-22 15:25:42 +00:00
Dan Gohman
5863f1e8c6 Add a comment explaining a non-obvious API.
llvm-svn: 106539
2010-06-22 13:57:08 +00:00
Devang Patel
f50b6147ad Use single interface, using twine, to get named metadata.
getNamedMetadata().

llvm-svn: 106518
2010-06-22 01:19:38 +00:00
Evan Cheng
a1ebf91a39 Tail merging pass shall not break up IT blocks. rdar://8115404
llvm-svn: 106517
2010-06-22 01:18:16 +00:00
Eric Christopher
68b77d45c2 Remove isTwoAddress from llvm.
llvm-svn: 106470
2010-06-21 20:35:09 +00:00
Dan Gohman
745765d0b8 Move several non-performance-critical member functinos out of line.
llvm-svn: 106444
2010-06-21 18:46:45 +00:00
Chris Lattner
42e833c9c4 make the Value constructor protected.
llvm-svn: 106427
2010-06-21 16:46:37 +00:00
Dan Gohman
fb091c213b Factor out duplicated code for reusing and inserting casts into
a helper function.

llvm-svn: 106388
2010-06-19 13:25:23 +00:00
Dan Gohman
823dff64cd Teach regular and fast isel to set dead flags on unused implicit defs
on calls and similar instructions.

llvm-svn: 106353
2010-06-18 23:28:01 +00:00
Evan Cheng
b5fadc47e0 Allow ARM if-converter to be run after post allocation scheduling.
- This fixed a number of bugs in if-converter, tail merging, and post-allocation
  scheduler. If-converter now runs branch folding / tail merging first to
  maximize if-conversion opportunities.
- Also changed the t2IT instruction slightly. It now defines the ITSTATE
  register which is read by instructions in the IT block.
- Added Thumb2 specific hazard recognizer to ensure the scheduler doesn't
  change the instruction ordering in the IT block (since IT mask has been
  finalized). It also ensures no other instructions can be scheduled between
  instructions in the IT block.

This is not yet enabled.

llvm-svn: 106344
2010-06-18 23:09:54 +00:00
Jim Grosbach
b8c94667a8 back-end libcall handling for ATOMIC_SWAP (__sync_lock_test_and_set)
llvm-svn: 106342
2010-06-18 23:03:10 +00:00
Jim Grosbach
91aae1c534 Add Expand-to-libcall support for additional atomics. This covers the usual
entries used by llvm-gcc. *_[U]MIN and such can be added later if needed.

This enables the front ends to simplify handling of the atomic intrinsics by
removing the target-specific decision about which targets can handle the
intrinsics.

llvm-svn: 106321
2010-06-18 21:43:38 +00:00
Dan Gohman
b5ec637e57 Revert r106304 (105548 and friends), which are the SCEVComplexityCompare
optimizations. There is still some nondeterminism remaining.

llvm-svn: 106306
2010-06-18 19:54:20 +00:00
Dan Gohman
527b570925 Reapply 105540, 105542, and 105548, and revert r105732.
llvm-svn: 106304
2010-06-18 19:26:04 +00:00
Dan Gohman
b220af45eb Add explicit keywords.
llvm-svn: 106300
2010-06-18 19:04:37 +00:00
Dan Gohman
1ccf40774e Start TargetRegisterClass indices at 0 instead of 1, so that
MachineRegisterInfo doesn't have to confusingly allocate an extra
entry.

llvm-svn: 106296
2010-06-18 18:13:55 +00:00
Jim Grosbach
60a3287950 Grammar.
llvm-svn: 106292
2010-06-18 17:40:42 +00:00
Dan Gohman
0422fad3a6 Don't bother calling releaseMemory before destroying the DominatorTreeBase.
llvm-svn: 106287
2010-06-18 16:09:11 +00:00
Dan Gohman
6c6e9a45ff Give NamedRegionTimer an Enabled flag, allowing all its clients to
switch from this:

  if (TimePassesIsEnabled) {
    NamedRegionTimer T(Name, GroupName);
    do_something();
  } else {
    do_something(); // duplicate the code, this time without a timer!
  }

to this:

  {
    NamedRegionTimer T(Name, GroupName, TimePassesIsEnabled);
    do_something();
  }

llvm-svn: 106285
2010-06-18 15:56:31 +00:00
Dan Gohman
e9dfb84007 Change UpdateNodeOperands' operand and return value from SDValue to
SDNode *, since it doesn't care about the ResNo value.

llvm-svn: 106282
2010-06-18 15:30:29 +00:00
Dan Gohman
e20e668c7e Remove getIntegerSCEV; it's redundant with getConstant, and getConstant
is more consistent with the ConstantInt API.

llvm-svn: 106281
2010-06-18 14:33:50 +00:00
Dan Gohman
0b5b93b20d Eliminate unnecessary uses of getZExtValue().
llvm-svn: 106279
2010-06-18 14:22:04 +00:00
Dan Gohman
22ff84372b isValueValidForType can be a static member function.
llvm-svn: 106278
2010-06-18 14:01:07 +00:00
Dan Gohman
8185674354 Fold the ShrinkDemandedOps pass into the regular DAGCombiner pass,
which is faster, simpler, and less surprising.

llvm-svn: 106263
2010-06-18 01:05:21 +00:00