Commit Graph

16191 Commits

Author SHA1 Message Date
Hal Finkel
4af941de1c Refactor operation equivalence checking in BBVectorize by extending Instruction::isSameOperationAs.
Maintaining this kind of checking in different places is dangerous, extending
Instruction::isSameOperationAs consolidates this logic into one place. Here
I've added an optional flags parameter and two flags that are important for
vectorization: CompareIgnoringAlignment and CompareUsingScalarTypes.

llvm-svn: 159329
2012-06-28 05:42:26 +00:00
Bill Wendling
e8949ecfa6 Move lib/Analysis/DebugInfo.cpp to lib/VMCore/DebugInfo.cpp and
include/llvm/Analysis/DebugInfo.h to include/llvm/DebugInfo.h.

The reasoning is because the DebugInfo module is simply an interface to the
debug info MDNodes and has nothing to do with analysis.

llvm-svn: 159312
2012-06-28 00:05:13 +00:00
Jack Carter
dc0ebcb076 The ELF relocation record format is different for N64
which many Mips 64 ABIs use than for O64 which many 
if not all other target ABIs use.

Most architectures have the following 64 bit relocation record format:

  typedef struct
  {
    Elf64_Addr   r_offset; /* Address of reference */
    Elf64_Xword  r_info;   /* Symbol index and type of relocation */
  } Elf64_Rel;

  typedef struct
  {
    Elf64_Addr    r_offset;
    Elf64_Xword   r_info;
    Elf64_Sxword  r_addend;
  } Elf64_Rela;

Whereas N64 has the following format:

  typedef struct
  {
    Elf64_Addr    r_offset;/* Address of reference */
    Elf64_Word  r_sym;     /* Symbol index */
    Elf64_Byte  r_ssym;    /* Special symbol */
    Elf64_Byte  r_type3;   /* Relocation type */
    Elf64_Byte  r_type2;   /* Relocation type */
    Elf64_Byte  r_type;    /* Relocation type */
  } Elf64_Rel;

  typedef struct
  {
    Elf64_Addr    r_offset;/* Address of reference */
    Elf64_Word  r_sym;     /* Symbol index */
    Elf64_Byte  r_ssym;    /* Special symbol */
    Elf64_Byte  r_type3;   /* Relocation type */
    Elf64_Byte  r_type2;   /* Relocation type */
    Elf64_Byte  r_type;    /* Relocation type */
    Elf64_Sxword  r_addend;
  } Elf64_Rela;

The structure is the same size, but the r_info data element 
is now 5 separate elements. Besides the content aspects, 
endian byte reordering will be different for the area with 
each element being endianized separately.

I treat this as generic and continue to pass r_type as 
an integer masking and unmasking the byte sized N64 
values for N64 mode. I've implemented this and it causes no 
affect on other current targets.

This passes make check.

Jack

llvm-svn: 159299
2012-06-27 22:28:30 +00:00
Matt Beaumont-Gay
93c66a3db1 Revert r159136 due to PR13124.
Original commit message:

If a constant or a function has linkonce_odr linkage and unnamed_addr, mark it
hidden. Being linkonce_odr guarantees that it is available in every dso that
needs it. Being a constant/function with unnamed_addr guarantees that the
copies don't have to be merged.

llvm-svn: 159272
2012-06-27 17:10:33 +00:00
Bill Wendling
ecfe03120d Revamp how debugging information is emitted for debug info objects.
It's not necessary for each DI class to have its own copy of `print' and
`dump'. Instead, just give DIDescriptor those methods and have it call the
appropriate debugging printing routine based on the type of the debug
information.

llvm-svn: 159237
2012-06-26 22:57:33 +00:00
Manman Ren
6be46b7b4c X86: add GATHER intrinsics (AVX2) in LLVM
Support the following intrinsics:
llvm.x86.avx2.gather.d.pd, llvm.x86.avx2.gather.q.pd
llvm.x86.avx2.gather.d.pd.256, llvm.x86.avx2.gather.q.pd.256
llvm.x86.avx2.gather.d.ps, llvm.x86.avx2.gather.q.ps
llvm.x86.avx2.gather.d.ps.256, llvm.x86.avx2.gather.q.ps.256

Modified Disassembler to handle VSIB addressing mode.

llvm-svn: 159221
2012-06-26 19:47:59 +00:00
Jakob Stoklund Olesen
ee1b7565a6 Allow targets to inject passes before the virtual register rewriter.
Such passes can be used to tweak the register assignments in a
target-dependent way, for example to avoid write-after-write
dependencies.

llvm-svn: 159209
2012-06-26 17:09:29 +00:00
Stepan Dyatkovskiy
ec5607e71d IntegersSubsetMapping: implemented "diff" operation. Operation allows at the same time perform up to three operations:
- LHS exclude RHS
- LHS intersect RHS (LHS successors will keeped)
- RHS exclude LHS
The complexity is N+M, where
  N is size of LHS
  M is size of RHS.

llvm-svn: 159201
2012-06-26 11:57:43 +00:00
Stepan Dyatkovskiy
612afe85e7 IntegersSubsetMapping: removed exclude operation, it will replaced with more universal "diff" operation in next commit.
Changes was separated onto two commits for better readability.

llvm-svn: 159200
2012-06-26 11:41:47 +00:00
Andrew Trick
c5e08120a4 Enable the new LoopInfo algorithm by default.
The primary advantage is that loop optimizations will be applied in a
stable order. This helps debugging and unit test creation. It is also
a better overall implementation without pathologically bad performance
on deep functions.

On large functions (llvm-stress --size=200000 | opt -loops)
Before: 0.1263s
After:  0.0225s

On deep functions (after tweaking llvm-stress, thanks Nadav):
Before: 0.2281s
After:  0.0227s

See r158790 for more comments.

The loop tree is now consistently generated in forward order, but loop
passes are applied in reverse order over the program. If we have a
loop optimization that prefers forward order, that can easily be
achieved by adding a different type of LoopPassManager.

llvm-svn: 159183
2012-06-26 04:11:38 +00:00
Owen Anderson
9dce23492c Define DAGOperand, an empty base class for RegisterClass and Operand. This allows one to write multiclasses that are polymorphic over both registers and non-register operands.
llvm-svn: 159162
2012-06-25 21:25:16 +00:00
Nuno Lopes
6cd9a63eb4 add CallSite/CallInst/InvokeInst::hasFnAttr()
llvm-svn: 159144
2012-06-25 16:16:58 +00:00
Rafael Espindola
45a2b18594 If a constant or a function has linkonce_odr linkage and unnamed_addr, mark it
hidden. Being linkonce_odr guarantees that it is available in every dso that
needs it. Being a constant/function with unnamed_addr guarantees that the
copies don't have to be merged.

llvm-svn: 159136
2012-06-25 14:30:31 +00:00
Eli Bendersky
5d45af3f75 The name (and comment describing) of llvm::GetFirstDebuigLocInBasicBlock no longer represents what the function does. Therefore, the function is removed and its functionality is folded into the only place in the code-base where it was being used.
llvm-svn: 159133
2012-06-25 10:13:14 +00:00
Chandler Carruth
4aeccfe390 Just remove generic support for C++11 alignas -- GCC is already
advertising complete support w/o alignas implemented, and its
implementation of alignas in the latest versions is so convoluted as to
be unusable.

llvm-svn: 159125
2012-06-25 05:20:13 +00:00
Hal Finkel
409cab2a0a Allow controlling vectorization of boolean values separately from other integer types.
These are used as the result of comparisons, and often handled differently from larger integer types.

llvm-svn: 159111
2012-06-24 13:28:01 +00:00
NAKAMURA Takumi
3da73accf9 llvm/Support/IntegersSubset.h: Add a copy constructor on IntegersSubset to appease msvc.
msvc mis-infers ParentTy(RHS) to (const RangesCollectionTy &).

llvm-svn: 159101
2012-06-24 03:48:53 +00:00
NAKAMURA Takumi
a7465f2530 llvm/Support/IntegersSubset.h: Fix whitespace.
llvm-svn: 159100
2012-06-24 03:48:47 +00:00
Hal Finkel
d0a65988d8 Allow BBVectorize to fuse compare instructions.
llvm-svn: 159088
2012-06-23 21:52:50 +00:00
Marshall Clow
b34fa3efd3 Add relocation types for Hexagon processor; patch by Sidney Manning <sidneym@codeaurora.org>
llvm-svn: 159081
2012-06-23 14:46:18 +00:00
Hans Wennborg
bdb7ec3dc8 Clean-up after r159077.
Remove temporary GlobalVariable constructors now that Clang has been
updated (r159078).

llvm-svn: 159079
2012-06-23 12:14:23 +00:00
Hans Wennborg
8c011bd43a Extend the IL for selecting TLS models (PR9788)
This allows the user/front-end to specify a model that is better
than what LLVM would choose by default. For example, a variable
might be declared as

  @x = thread_local(initialexec) global i32 42

if it will not be used in a shared library that is dlopen'ed.

If the specified model isn't supported by the target, or if LLVM can
make a better choice, a different model may be used.

llvm-svn: 159077
2012-06-23 11:37:03 +00:00
Stepan Dyatkovskiy
bc2affc34a Optimized usage of new SwitchInst case values (IntegersSubset type) in Local.cpp, Execution.cpp and BitcodeWriter.cpp.
I got about 1% of compile-time improvement on my machines (Ubuntu 11.10 i386 and Ubuntu 12.04 x64).

llvm-svn: 159076
2012-06-23 10:58:58 +00:00
Jim Grosbach
9bfc8e18b0 TableGen: AsmMatcher support for better operand diagnostics.
"Invalid operand" may be a completely correct diagnostic, but it's often
insufficiently specific to really help identify and fix the problem in
assembly source. Allow a target to specify a more-specific diagnostic kind
for each AsmOperandClass derived definition and use that to provide
more detailed diagnostics when an operant of that class resulted in a
match failure.

rdar://8987109

llvm-svn: 159050
2012-06-22 23:56:44 +00:00
Jakob Stoklund Olesen
70a37b6a67 Remove ProcessImplicitDefs.h which was unused.
The ProcessImplicitDefs class can be local to its implementation file.

llvm-svn: 159041
2012-06-22 22:27:36 +00:00
Jakob Stoklund Olesen
3a972a4f8d Delete a boring statistic.
llvm-svn: 159030
2012-06-22 20:40:15 +00:00
Jakob Stoklund Olesen
5b5a4305f1 Store live intervals in an IndexedMap.
It is both smaller and faster than DenseMap.

llvm-svn: 159029
2012-06-22 20:37:52 +00:00
Hal Finkel
db4f1462bf Revert r158679 - use case is unclear (and it increases the memory footprint).
Original commit message:
    Allow up to 64 functional units per processor itinerary.

    This patch changes the type used to hold the FU bitset from unsigned to uint64_t.
    This will be needed for some upcoming PowerPC itineraries.

llvm-svn: 159027
2012-06-22 20:27:13 +00:00
Evan Cheng
d957460992 EmitZerofill should take a 64-bit size or else it's chopping off large zero-filled global. rdar://11729134
llvm-svn: 159023
2012-06-22 20:14:46 +00:00
Jakob Stoklund Olesen
a925ef2596 Stop computing physreg live ranges.
Everyone is using on-demand regunit ranges now.

llvm-svn: 159018
2012-06-22 18:20:50 +00:00
Kaelyn Uhrain
47de25f313 Remove a variable that is unused when assertions aren't enabled.
llvm-svn: 159011
2012-06-22 17:18:15 +00:00
Jakob Stoklund Olesen
0d48b013fb Remove LiveIntervals::trackingRegUnits().
With regunit liveness permanently enabled, this function would always
return true.

Also remove now obsolete code for checking physreg interference.

llvm-svn: 159006
2012-06-22 16:46:44 +00:00
Dmitri Gribenko
1d11634bbe Change comment into proper Doxygen member comment.
llvm-svn: 159000
2012-06-22 16:00:48 +00:00
Stepan Dyatkovskiy
2e9bfd1dfc Fixed r158979.
Original message:
Performance optimizations:
- SwitchInst: case values stored separately from Operands List. It allows to make faster access to individual case value numbers or ranges.
- Optimized IntItem, added APInt value caching.
- Optimized IntegersSubsetGeneric: added optimizations for cases when subset is single number or when subset consists from single numbers only.

llvm-svn: 158997
2012-06-22 14:53:30 +00:00
Rafael Espindola
0280a5d85b Remove another duplicated variable. We only need one to tell us if the linker
knows dwarf or not.

llvm-svn: 158993
2012-06-22 13:32:49 +00:00
Rafael Espindola
13084dd6a3 Fix a FIXME: DwarfRequiresRelocationForSectionOffset is the same as
DwarfUsesRelocationsAcrossSections.

llvm-svn: 158992
2012-06-22 13:24:07 +00:00
Duncan Sands
2683b91b34 Revert commit 158979 (dyatkovskiy) since it is causing several buildbots to
fail.  Original commit message:

Performance optimizations:
- SwitchInst: case values stored separately from Operands List. It allows to make faster access to individual case value numbers or ranges.
- Optimized IntItem, added APInt value caching.
- Optimized IntegersSubsetGeneric: added optimizations for cases when subset is single number or when subset consists from single numbers only.

On my machine these optimizations gave about 4-6% of compile-time improvement.

llvm-svn: 158986
2012-06-22 10:35:06 +00:00
Stepan Dyatkovskiy
030f6bfdde Performance optimizations:
- SwitchInst: case values stored separately from Operands List. It allows to make faster access to individual case value numbers or ranges.
- Optimized IntItem, added APInt value caching.
- Optimized IntegersSubsetGeneric: added optimizations for cases when subset is single number or when subset consists from single numbers only.

On my machine these optimizations gave about 4-6% of compile-time improvement.

llvm-svn: 158979
2012-06-22 07:35:13 +00:00
Andrew Trick
279bd30bbc Use "NoItineraries" for processors with no itineraries.
This makes it explicit when ScoreboardHazardRecognizer will be used.
"GenericItineraries" would only make sense if it contained real
itinerary values and still required ScoreboardHazardRecognizer.

llvm-svn: 158963
2012-06-22 03:58:51 +00:00
Nick Lewycky
da52706728 Emit relocations for DW_AT_location entries on systems which need it. This is
a recommit of r127757. Fixes PR9493. Patch by Paul Robinson!

llvm-svn: 158957
2012-06-22 01:25:12 +00:00
Lang Hames
68cf87e3ef Rename -allow-excess-fp-precision flag to -fuse-fp-ops, and switch from a
boolean flag to an enum: { Fast, Standard, Strict } (default = Standard).

This option controls the creation by optimizations of fused FP ops that store
intermediate results in higher precision than IEEE allows (E.g. FMAs). The
behavior of this option is intended to match the behaviour specified by a
soon-to-be-introduced frontend flag: '-ffuse-fp-ops'.

Fast mode - allows formation of fused FP ops whenever they're profitable.

Standard mode - allow fusion only for 'blessed' FP ops. At present the only
blessed op is the fmuladd intrinsic. In the future more blessed ops may be
added.

Strict mode - allow fusion only if/when it can be proven that the excess
precision won't effect the result.

Note: This option only controls formation of fused ops by the optimizers.  Fused
operations that are explicitly requested (e.g. FMA via the llvm.fma.* intrinsic)
will always be honored, regardless of the value of this option.

Internally TargetOptions::AllowExcessFPPrecision has been replaced by
TargetOptions::AllowFPOpFusion.

llvm-svn: 158956
2012-06-22 01:09:09 +00:00
Nuno Lopes
1393c03048 remove extractMallocCallFromBitCast, since it was tailor maded for its sole user. Update GlobalOpt accordingly.
llvm-svn: 158952
2012-06-22 00:25:01 +00:00
Nuno Lopes
8baf9fdf84 Add support for invoke to the MemoryBuiltin analysid.
Update comments accordingly.

Make instcombine remove useless invokes to C++'s 'new' allocation function (test attached).

llvm-svn: 158937
2012-06-21 21:25:05 +00:00
Nuno Lopes
3d029880cc move some typedefs so that we don't polute the llvm namespace. this should appease the GCC buildbots
llvm-svn: 158924
2012-06-21 16:58:41 +00:00
Nuno Lopes
c9edab11db refactor the MemoryBuiltin analysis:
- provide more extensive set of functions to detect library allocation functions (e.g., malloc, calloc, strdup, etc)
 - provide an API to compute the size and offset of an object pointed by

Move a few clients (GVN, AA, instcombine, ...) to the new API.
This implementation is a lot more aggressive than each of the custom implementations being replaced.

Patch reviewed by Nick Lewycky and Chandler Carruth, thanks.

llvm-svn: 158919
2012-06-21 15:45:28 +00:00
Nadav Rotem
313b090606 Add a number of threshold arguments to the SRA pass.
A patch by Tom Stellard with minor changes.

llvm-svn: 158918
2012-06-21 13:44:31 +00:00
Jakob Stoklund Olesen
ee49ea0940 Remove LiveIntervals::iterator.
Live intervals for regunits and virtual registers are stored separately,
and physreg live intervals are going away.

To visit the live ranges of all virtual registers, use this pattern
instead:

  for (unsigned i = 0, e = MRI->getNumVirtRegs(); i != e; ++i) {
    unsigned Reg = TargetRegisterInfo::index2VirtReg(i);
    if (MRI->reg_nodbg_empty(Reg))
      continue;

llvm-svn: 158879
2012-06-20 23:54:20 +00:00
Jakob Stoklund Olesen
75fa4ceaa1 Remove the RenderMachineFunction HTML output pass.
I don't think anyone has been using this functionality for a while, and
it is getting in the way of refactoring now.

llvm-svn: 158876
2012-06-20 23:47:58 +00:00
Andrew Trick
33bf2f7893 Restructure PopulateLoopsDFS::insertIntoLoop.
As Nadav pointed out the first implementation was obscure.

llvm-svn: 158862
2012-06-20 22:18:33 +00:00
Andrew Trick
8b17d495d7 Add "extern template" declarations now that we use explicit instantiation.
This is supported by gcc and clang, but guarded by a macro for MSVC 2008.

The extern template declaration is not necessary but generally good
form. It can avoid extra instantiations of the template methods
defined inline.

The EXTERN_TEMPLATE_INSTANTIATION macro could probably be generalized to
handle multiple template parameters if someone thinks it's worthwhile.

llvm-svn: 158840
2012-06-20 20:17:20 +00:00
Jakob Stoklund Olesen
6d2db5c3d9 Only update regunit live ranges that have been precomputed.
Regunit live ranges are computed on demand, so when mi-sched calls
handleMove, some regunits may not have live ranges yet.

That makes updating them easier: Just skip the non-existing ranges. They
will be computed correctly from the rescheduled machine code when they
are needed.

llvm-svn: 158831
2012-06-20 18:00:57 +00:00
Chandler Carruth
6f8cc37074 Remove 'static' from inline functions defined in header files.
There is a pretty staggering amount of this in LLVM's header files, this
is not all of the instances I'm afraid. These include all of the
functions that (in my build) are used by a non-static inline (or
external) function. Specifically, these issues were caught by the new
'-Winternal-linkage-in-inline' warning.

I'll try to just clean up the remainder of the clearly redundant "static
inline" cases on functions (not methods!) defined within headers if
I can do so in a reliable way.

There were even several cases of a missing 'inline' altogether, or my
personal favorite "static bool inline". Go figure. ;]

llvm-svn: 158800
2012-06-20 08:39:33 +00:00
Andrew Trick
55aafa3d2e A new algorithm for computing LoopInfo. Temporarily disabled.
-stable-loops enables a new algorithm for generating the Loop
forest. It differs from the original algorithm in a few respects:

- Not determined by use-list order.
- Initially guarantees RPO order of block and subloops.
- Linear in the number of CFG edges.
- Nonrecursive.

I didn't want to change the LoopInfo API yet, so the block lists are
still inclusive. This seems strange to me, and it means that building
LoopInfo is not strictly linear, but it may not be a problem in
practice. At least the block lists start out in RPO order now. In the
future we may add an attribute or wrapper analysis that allows other
passes to assume RPO order.

The primary motivation of this work was not to optimize LoopInfo, but
to allow reproducing performance issues by decomposing the compilation
stages. I'm often unable to do this with the current LoopInfo, because
the loop tree order determines Loop pass order. Serializing the IR
tends to invert the order, which reverses the optimization order. This
makes it nearly impossible to debug interdependent loop optimizations
such as LSR.

I also believe this will provide more stable performance results across time.

llvm-svn: 158790
2012-06-20 05:23:33 +00:00
Andrew Trick
0db6403960 Move the implementation of LoopInfo into LoopInfoImpl.h.
The implementation only needs inclusion from LoopInfo.cpp and
MachineLoopInfo.cpp. Clients of the interface should only include the
interface. This makes the interface readable and speeds up rebuilds
after modifying the implementation.

llvm-svn: 158787
2012-06-20 03:42:09 +00:00
Nick Kledzik
d4f7f6f97a Add permissions(), map_file_pages(), and unmap_file_pages() to llvm::sys::fs and add unit test. Unix is implemented. Windows side needs to be implemented.
llvm-svn: 158770
2012-06-20 00:28:54 +00:00
Chad Rosier
76e60fbe46 Add an ensureMaxAlignment() function to MachineFrameInfo (analogous to
ensureAlignment() in MachineFunction).  Also, drop setMaxAlignment() in
favor of this new function.  This creates a main entry point to setting
MaxAlignment, which will be helpful for future work.  No functionality
change intended.

llvm-svn: 158758
2012-06-19 22:59:12 +00:00
Lang Hames
f0b9601a6d Add DAG-combines for aggressive FMA formation.
This patch adds DAG combines to form FMAs from pairs of FADD + FMUL or
FSUB + FMUL. The combines are performed when:
(a) Either
      AllowExcessFPPrecision option (-enable-excess-fp-precision for llc)
        OR
      UnsafeFPMath option (-enable-unsafe-fp-math)
    are set, and
(b) TargetLoweringInfo::isFMAFasterThanMulAndAdd(VT) is true for the type of
    the FADD/FSUB, and
(c) The FMUL only has one user (the FADD/FSUB).

If your target has fast FMA instructions you can make use of these combines by
overriding TargetLoweringInfo::isFMAFasterThanMulAndAdd(VT) to return true for
types supported by your FMA instruction, and adding patterns to match ISD::FMA
to your FMA instructions.

llvm-svn: 158757
2012-06-19 22:51:23 +00:00
Chad Rosier
0864fce514 Typo. Patch by Cameron McInally <cameron.mcinally@nyu.edu>.
llvm-svn: 158754
2012-06-19 22:28:18 +00:00
Rafael Espindola
38c45a939d Move the support for using .init_array from ARM to the generic
TargetLoweringObjectFileELF. Use this to support it on X86. Unlike ARM,
on X86 it is not easy to find out if .init_array should be used or not, so
the decision is made via TargetOptions and defaults to off.

Add a command line option to llc that enables it.

llvm-svn: 158692
2012-06-19 00:48:28 +00:00
Nuno Lopes
de7b3a54f2 revert r158660, since Chris has some issues with this patch (namely using code to reprent information only used by the compiler)
Original commit msg:
add the 'alloc' metadata node to represent the size of offset of buffers pointed to by pointers.
This metadata can be attached to any instruction returning a pointer

llvm-svn: 158688
2012-06-18 23:34:26 +00:00
David Blaikie
9205346d93 Don't copy a potentially-uninitialized variable.
Based on review discussion of r158638 with Chandler Carruth, Tobias von Koch, and Duncan Sands and a -Wmaybe-uninitialized warning from GCC.

llvm-svn: 158685
2012-06-18 22:31:28 +00:00
Hal Finkel
56f4d93767 Allow up to 64 functional units per processor itinerary.
This patch changes the type used to hold the FU bitset from unsigned to uint64_t.
This will be needed for some upcoming PowerPC itineraries.

llvm-svn: 158679
2012-06-18 21:08:18 +00:00
Marshall Clow
c56632d5e5 Added accessors for getting coff_relocation info
llvm-svn: 158675
2012-06-18 19:47:16 +00:00
Nuno Lopes
aa5ffcb407 add the 'alloc' metadata node to represent the size of offset of buffers pointed to by pointers.
This metadata can be attached to any instruction returning a pointer

llvm-svn: 158660
2012-06-18 16:04:04 +00:00
Benjamin Kramer
d408a3b134 Bring the return value of SmallVector::insert in line with std::vector::insert.
It always returns the iterator for the first inserted element, or the passed in
iterator if the inserted range was empty. Flesh out the unit test more and fix
all the cases it uncovered so far.

llvm-svn: 158645
2012-06-17 12:46:13 +00:00
Chandler Carruth
f0352b4be5 Remove SmallMap, and the several files that were used to implement it.
We have SmallDenseMap now that has more correct and predictable
semantics, even though it is a more narrow abstraction.

llvm-svn: 158644
2012-06-17 12:07:42 +00:00
Benjamin Kramer
f78d0a14b0 SmallVector: return a valid iterator for the rare case of inserting an empty range into a SmallVector.
Patch by Johannes Schaub!

llvm-svn: 158643
2012-06-17 11:52:22 +00:00
Chandler Carruth
5b13784777 Add a unit test for 'swap', and fix a pile of bugs in
SmallDenseMap::swap.

First, make it parse cleanly. Yay for uninstantiated methods.

Second, make the inline-buckets case work correctly. This is way
trickier than it should be due to the uninitialized values in empty and
tombstone buckets.

Finally fix a few typos that caused construction/destruction mismatches
in the counting unittest.

llvm-svn: 158641
2012-06-17 11:28:13 +00:00
Chandler Carruth
ea5804adca Add tests for *DenesMap for both key and value types' construction and
destruction and fix a bug in SmallDenseMap they caught.

This is kind of a poor-man's version of the testing that just adds the
addresses to a set on construction and removes them on destruction. We
check that double construction and double destruction don't occur.
Amusingly enough, this is enough to catch a lot of SmallDenseMap issues
because we spend a lot of time with fixed stable addresses in the inline
buffer.

The SmallDenseMap bug fix included makes grow() not double-destroy in
some cases. It also fixes a FIXME there, the code was pretty crappy. We
now don't have any wasted initialization, but we do move the entries in
inline bucket array an extra time. It's probably a better tradeoff, and
is much easier to get correct.

llvm-svn: 158639
2012-06-17 10:33:51 +00:00
Chandler Carruth
1501b6be37 Introduce a SmallDenseMap container that re-uses the existing DenseMap
implementation.

This type includes an inline bucket array which is used initially. Once
it is exceeded, an array of 64 buckets is allocated on the heap. The
bucket count grows from there as needed. Some highlights of this
implementation:

- The inline buffer is very carefully aligned, and so supports types
  with alignment constraints.
- It works hard to avoid aliasing issues.
- Supports types with non-trivial constructors, destructors, copy
  constructions, etc. It works reasonably hard to minimize copies and
  unnecessary initialization. The most common initialization is to set
  keys to the empty key, and so that should be fast if at all possible.

This class has a performance / space trade-off. It tries to optimize for
relatively small maps, and so packs the inline bucket array densely into
the object. It will be marginally slower than a normal DenseMap in a few
use patterns, so it isn't appropriate everywhere.

The unit tests for DenseMap have been generalized a bit to support
running over different map implementations in addition to different
key/value types. They've then been automatically extended to cover the
new container through the magic of GoogleTest's typed tests.

All of this is still a bit rough though. I'm going to be cleaning up
some aspects of the implementation, documenting things better, and
adding tests which include non-trivial types. As soon as I'm comfortable
with the correctness, I plan to switch existing users of SmallMap over
to this class as it is already more correct w.r.t. construction and
destruction of objects iin the map.

Thanks to Benjamin Kramer for all the reviews of this and the lead-up
patches. That said, more review on this would really be appreciated. As
I've noted a few times, I'm quite surprised how hard it is to get the
semantics for a hashtable-based map container with a small buffer
optimization correct. =]

llvm-svn: 158638
2012-06-17 09:05:09 +00:00
Benjamin Kramer
cd70a90dd5 Guard private fields that are unused in Release builds with #ifndef NDEBUG.
llvm-svn: 158608
2012-06-16 21:48:13 +00:00
Hal Finkel
66e13debff Move the Metadata merging methods from GVN and make them public in MDNode.
There are other passes, BBVectorize specifically, that also need some of
this functionality.

llvm-svn: 158605
2012-06-16 20:33:37 +00:00
Benjamin Kramer
ed052699c6 Merge the SmallBitVector and BitVector unit tests with gtest's typed test magic and bring SmallBitVector up to date.
llvm-svn: 158600
2012-06-16 10:51:07 +00:00
Chandler Carruth
7b9797e637 Add support to the alignment support header for conjuring a character
array of a suitable size and alignment for any of a number of different
types to be stored into the character array.

The mechanisms for producing an explicitly aligned type are fairly
complex because this operation is poorly supported on all compilers.
We've spent a fairly significant amount of time experimenting with
different implementations inside of Google, and the one using explicitly
expanded templates has been the most robust.

Credit goes to Nick Lewycky for writing the first 20 versions or so of
this logic we had inside of Google. I based this on the only one to
actually survive. In case anyone is worried, yes we are both explicitly
re-contributing and re-licensing it for LLVM. =]

Once the issues with actually specifying the alignment are finished, it
turns out that most compilers don't in turn align anything the way they
are instructed. Testing of this logic against both Clang and GCC
indicate that the alignment constraints are largely ignored by both
compilers! I've come up with and used a work-around by wrapping each
alignment-hinted type directly in a struct, and using that struct to
align the character array through a union. This elaborate hackery is
terrifying, but I've included testing that caught a terrifying number of
bugs in every other technique I've tried.

All of this in order to implement a poor C++98 programmers emulation of
C++11 unrestricted unions in classes such as SmallDenseMap.

llvm-svn: 158597
2012-06-16 08:52:57 +00:00
Chandler Carruth
bdeac17070 Lift the NumElements and NumTombstones members into the super class
rather than the base class. Add a pile of boilerplate to indirect around
this.

This is pretty ugly, but it allows the super class to change the
representation of these values, which will be key for doing
a SmallDenseMap.

Suggestions on better method structuring / naming are welcome, but keep
in mind that SmallDenseMap won't have an 'unsigned' member to expose
a reference to... =/

llvm-svn: 158586
2012-06-16 01:18:07 +00:00
Chandler Carruth
d3d541a9d2 Factor DenseMap into a base class that implements the hashtable logic,
and a derived class that provides the allocation and growth strategy.

This is the first (and biggest) step toward building a SmallDenseMap
that actually behaves exactly the same as DenseMap, and supports all the
same types and interface points with the same semantics.

llvm-svn: 158585
2012-06-16 01:05:01 +00:00
Marshall Clow
49e964c47f Adding acessors to COFFObjectFile so that clients can get at the (non-generic) bits
llvm-svn: 158484
2012-06-15 01:08:25 +00:00
Rafael Espindola
41f1be2080 Implement the isSafeToDiscardIfUnused predicate and use it in globalopt and
globaldce. Globaldce was already removing linkonce globals, but globalopt was
not.

llvm-svn: 158476
2012-06-14 22:48:13 +00:00
Stepan Dyatkovskiy
bb6d2dc9da SmallMap, FlatArrayMap::copyFrom
Replaced memcpy with std::copy, since the first one may work improperly with non POD data.

llvm-svn: 158457
2012-06-14 16:59:43 +00:00
Chandler Carruth
e63fe55776 Group the 'unsigned' members after the pointer to avoid 4 bytes of
padding on x86-64.

llvm-svn: 158421
2012-06-13 21:44:07 +00:00
Kay Tiong Khoo
b631f7fd59 *typo: Cyles changed to Cycles
llvm-svn: 158404
2012-06-13 15:53:04 +00:00
Duncan Sands
5f04c03e66 When linearizing a multiplication, return at once if we see a factor of zero,
since then the entire expression must equal zero (similarly for other operations
with an absorbing element).  With this in place a bunch of reassociate code for
handling constants is dead since it is all taken care of when linearizing.  No
intended functionality change.

llvm-svn: 158398
2012-06-13 09:42:13 +00:00
Craig Topper
b2299168d3 Fix intrinsics for XOP frczss/sd instructions. These instructions only take one source register and zero the upper bits of the destination rather than preserving them.
llvm-svn: 158396
2012-06-13 07:18:53 +00:00
Jakob Stoklund Olesen
3815bb6f62 Eliminate struct TableGenBackend.
TableGen backends are simply written as functions now.

Patch by Sean Silva!

llvm-svn: 158389
2012-06-13 05:15:49 +00:00
Andrew Trick
858cf20add sched: Avoid trivially redundant DAG edges. Take the one with higher latency.
llvm-svn: 158379
2012-06-13 02:39:00 +00:00
David Blaikie
87a8e6af9f Remove use of GNU extension to resolve Clang warning.
llvm-svn: 158364
2012-06-12 17:06:32 +00:00
Duncan Sands
5948d230e5 Now that Reassociate's LinearizeExprTree can look through arbitrary expression
topologies, it is quite possible for a leaf node to have huge multiplicity, for
example: x0 = x*x, x1 = x0*x0, x2 = x1*x1, ... rapidly gives a value which is x
raised to a vast power (the multiplicity, or weight, of x).  This patch fixes
the computation of weights by correctly computing them no matter how big they
are, rather than just overflowing and getting a wrong value.  It turns out that
the weight for a value never needs more bits to represent than the value itself,
so it is enough to represent weights as APInts of the same bitwidth and do the
right overflow-avoiding dance steps when computing weights.  As a side-effect it
reduces the number of multiplies needed in some cases of large powers.  While
there, in view of external uses (eg by the vectorizer) I made LinearizeExprTree
static, pushing the rank computation out into users.  This is progress towards
fixing PR13021.

llvm-svn: 158358
2012-06-12 14:33:56 +00:00
Argyrios Kyrtzidis
30b183277e Satisfy C++ aliasing rules, per suggestion by Chandler.
llvm-svn: 158346
2012-06-12 01:06:16 +00:00
Argyrios Kyrtzidis
c10e1df8d8 For llvm::sys::ThreadLocalImpl instead of malloc'ing the platform-specific
thread local data, embed them in the class using a uint64_t and make sure
we get compiler errors if there's a platform where this is not big enough.

This makes ThreadLocal more safe for using it in conjunction with CrashRecoveryContext.

Related to crash in rdar://11434201.

llvm-svn: 158342
2012-06-12 00:21:31 +00:00
Andrew Trick
642c7264bd misched: When querying RegisterPressureTracker, always save current and max pressure.
llvm-svn: 158340
2012-06-11 23:42:23 +00:00
Jakob Stoklund Olesen
2e87ef0470 Write llvm-tblgen backends as functions instead of sub-classes.
The TableGenBackend base class doesn't do much, and will be removed
completely soon.

Patch by Sean Silva!

llvm-svn: 158311
2012-06-11 15:37:55 +00:00
Jakob Stoklund Olesen
58b05fbad5 Fix a problem with the reverse bundle iterators.
This showed up the first time rend() was called on a bundled instruction
in the Mips backend.

Also avoid dereferencing end() in bundle_iterator::operator++().

We still don't have a place to put unit tests for this stuff.

llvm-svn: 158310
2012-06-11 15:11:12 +00:00
Craig Topper
b355582afd Add intrinsics for immediate form of XOP vprot instructions. Use i128mem instead of f128mem for integer XOP instructions.
llvm-svn: 158291
2012-06-10 07:31:56 +00:00
Craig Topper
ad5e38e410 Replace XOP vpcom intrinsics with fewer intrinsics that take the immediate as an argument.
llvm-svn: 158278
2012-06-09 16:46:13 +00:00
Benjamin Kramer
d9f9d9d77d Hashing: Remove outdated comment. Support for reserved hash values was removed in r151865.
llvm-svn: 158276
2012-06-09 15:33:28 +00:00
Andrew Trick
e5a1b98d5d Register pressure: added getPressureAfterInstr.
llvm-svn: 158256
2012-06-09 02:16:58 +00:00
Jakob Stoklund Olesen
c590d6ca6d Sketch a LiveRegMatrix analysis pass.
The LiveRegMatrix represents the live range of assigned virtual
registers in a Live interval union per register unit. This is not
fundamentally different from the interference tracking in RegAllocBase
that both RABasic and RAGreedy use.

The important differences are:

- LiveRegMatrix tracks interference per register unit instead of per
  physical register. This makes interference checks cheaper and
  assignments slightly more expensive. For example, the ARM D7 reigster
  has 24 aliases, so we would check 24 physregs before assigning to one.
  With unit-based interference, we check 2 units before assigning to 2
  units.

- LiveRegMatrix caches regmask interference checks. That is currently
  duplicated functionality in RABasic and RAGreedy.

- LiveRegMatrix is a pass which makes it possible to insert
  target-dependent passes between register allocation and rewriting.
  Such passes could tweak the register assignments with interference
  checking support from LiveRegMatrix.

Eventually, RABasic and RAGreedy will be switched to LiveRegMatrix.

llvm-svn: 158255
2012-06-09 02:13:10 +00:00
Dmitri Gribenko
6319fd5eb8 Convert comments to proper Doxygen comments.
llvm-svn: 158248
2012-06-09 00:01:45 +00:00
Andrew Trick
a1df722f41 Removing strange "using" declarations form TargetInstrInfo.
I can't imagine why these were added. Trial and error.

llvm-svn: 158247
2012-06-08 23:56:26 +00:00
Jakob Stoklund Olesen
c0bb0e899d Reintroduce VirtRegRewriter.
OK, not really. We don't want to reintroduce the old rewriter hacks.

This patch extracts virtual register rewriting as a separate pass that
runs after the register allocator. This is possible now that
CodeGen/Passes.cpp can configure the full optimizing register allocator
pipeline.

The rewriter pass uses register assignments in VirtRegMap to rewrite
virtual registers to physical registers, and it inserts kill flags based
on live intervals.

These finalization steps are the same for the optimizing register
allocators: RABasic, RAGreedy, and PBQP.

llvm-svn: 158244
2012-06-08 23:44:45 +00:00
Andrew Trick
151209d9dc TargetInstrInfo hooks implemented in codegen should be declared pure virtual.
llvm-svn: 158233
2012-06-08 21:52:38 +00:00
Andrew Trick
678eee83ab Sched itinerary fix: Avoid static initializers.
This fixes an accidental dependence on static initialization order that I introduced yesterday.

Thank you Lang!!!

llvm-svn: 158215
2012-06-08 18:25:47 +00:00
Andrew Trick
4fe40f02fd Continue factoring computeOperandLatency. Use it for ARM hasHighOperandLatency.
llvm-svn: 158164
2012-06-07 19:42:04 +00:00
Pete Cooper
6a68b66241 Add internal read flags to MachineInstrBuilder and hook them into the MachineOperand flag of the same name
llvm-svn: 158137
2012-06-07 04:43:52 +00:00
Manman Ren
f591de61da Revert r157755.
The commit is intended to fix rdar://11540023.
It is implemented as part of peephole optimization. We can actually implement
this in the SelectionDAG lowering phase.

llvm-svn: 158122
2012-06-06 23:53:03 +00:00
Andrew Trick
3e809a2fba Move RegisterClassInfo.h.
Allow targets to access this API. It's required for RegisterPressure.

llvm-svn: 158102
2012-06-06 20:29:31 +00:00
Andrew Trick
c1ae96787a Move RegisterPressure.h.
Make it a general utility for use by Targets.

llvm-svn: 158097
2012-06-06 19:47:35 +00:00
Benjamin Kramer
58b98297ac Round 2 of dead private variable removal.
LLVM is now -Wunused-private-field clean except for
- lib/MC/MCDisassembler/Disassembler.h. Not sure why it keeps all those unaccessible fields.
- gtest.

llvm-svn: 158096
2012-06-06 19:47:08 +00:00
Benjamin Kramer
d93c18846c Remove unused private fields found by clang's new -Wunused-private-field.
There are some that I didn't remove this round because they looked like
obvious stubs. There are dead variables in gtest too, they should be
fixed upstream.

llvm-svn: 158090
2012-06-06 18:25:08 +00:00
Jakob Stoklund Olesen
12dac91467 Simplify LiveInterval::print().
Don't print out the register number and spill weight, making the TRI
argument unnecessary.

This allows callers to interpret the reg field. It can currently be a
virtual register, a physical register, a spill slot, or a register unit.

llvm-svn: 158031
2012-06-05 22:51:54 +00:00
Jakob Stoklund Olesen
f12252632a Add experimental support for register unit liveness.
Instead of computing a live interval per physreg, LiveIntervals can
compute live intervals per register unit. This makes impossible the
confusing situation where aliasing registers could have overlapping live
intervals. It should also make fixed interferernce checking cheaper
since registers have fewer register units than aliases.

Live intervals for regunits are computed on demand, using MRI use-def
chains and the new LiveRangeCalc class. Only regunits live in to ABI
blocks are precomputed during LiveIntervals::runOnMachineFunction().

The regunit liveness computations don't depend on LiveVariables.

llvm-svn: 158029
2012-06-05 22:02:15 +00:00
Jakob Stoklund Olesen
2ea0fd4346 Implement LiveRangeCalc::extendToUses() and createDeadDefs().
These LiveRangeCalc methods are to be used when computing a live range
from scratch.

llvm-svn: 158027
2012-06-05 21:54:09 +00:00
Andrew Trick
24cce40009 misched: API for minimum vs. expected latency.
Minimum latency determines per-cycle scheduling groups.
Expected latency determines critical path and cost.

llvm-svn: 158021
2012-06-05 21:11:27 +00:00
Lang Hames
30d8fb4deb Add a new intrinsic: llvm.fmuladd. This intrinsic represents a multiply-add
expression (a * b + c) that can be implemented as a fused multiply-add (fma)
if the target determines that this will be more efficient. This intrinsic
will be used to implement FP_CONTRACT support and an aggressive FMA formation
mode.

If your target has a fast FMA instruction you should override the
isFMAFasterThanMulAndAdd method in TargetLowering to return true.

llvm-svn: 158014
2012-06-05 19:07:46 +00:00
Jakob Stoklund Olesen
fa2f5b6a3f Remove dead function.
llvm-svn: 158005
2012-06-05 17:19:07 +00:00
Stepan Dyatkovskiy
37b3fe7fa7 IntegersSubsetMapping: added exclude operation, that allows to exclude subset of integers from current mapping.
llvm-svn: 157989
2012-06-05 07:57:36 +00:00
Stepan Dyatkovskiy
b0bae92021 IntegersSubsetMapping:
Changed type of Items collection: from std::vector to std::list.
Also some small fixes made in IntegersSubset.h, IntegersSubsetMapping.h and IntegersSubsetTest.cpp.

llvm-svn: 157987
2012-06-05 07:43:08 +00:00
Andrew Trick
e7159e6731 misched: Added MultiIssueItineraries.
This allows a subtarget to explicitly specify the issue width and
other properties without providing pipeline stage details for every
instruction.

llvm-svn: 157979
2012-06-05 03:44:40 +00:00
Andrew Trick
85bf12c584 misched: Allow disabling scoreboard hazard checking for subtargets with a
valid itinerary but no pipeline stages.

An itinerary can contain useful scheduling information without specifying pipeline stages for each instruction.

llvm-svn: 157977
2012-06-05 03:44:32 +00:00
Andrew Trick
8b333df134 whitespace
llvm-svn: 157976
2012-06-05 03:44:29 +00:00
Jakob Stoklund Olesen
4e37f536bf Remove the last remat-related code from LiveIntervalAnalysis.
Rematerialization is handled by LiveRangeEdit now.

llvm-svn: 157974
2012-06-05 01:06:15 +00:00
Jakob Stoklund Olesen
6e8c9ed65e Stop using LiveIntervals::isReMaterializable().
It is an old function that does a lot more than required by
CalcSpillWeights, which was the only remaining caller.

The isRematerializable() function never actually sets the isLoad
argument, so don't try to compute that.

llvm-svn: 157973
2012-06-05 01:06:12 +00:00
Jakob Stoklund Olesen
04457a5cc7 Delete dead code.
llvm-svn: 157963
2012-06-04 23:01:41 +00:00
Jakob Stoklund Olesen
17fcb85c11 Switch LiveIntervals member variable to LLVM naming standards.
No functional change.

llvm-svn: 157957
2012-06-04 22:39:14 +00:00
Roman Divacky
0daa2c0556 Implement local-exec TLS on PowerPC.
llvm-svn: 157935
2012-06-04 17:36:38 +00:00
Nadav Rotem
6969a673e9 Remove the "-promote-elements" flag. This flag is now enabled by default.
llvm-svn: 157925
2012-06-04 11:27:21 +00:00
Duncan Sands
2f70cb3b88 getAllOnesValue also works for vectors of integers.
llvm-svn: 157915
2012-06-04 07:18:12 +00:00
NAKAMURA Takumi
8b305bfcd3 IntRange: Restore the copy constuctor explicitly to appase buildbot.
llvm-svn: 157901
2012-06-03 15:42:12 +00:00
Craig Topper
8d3031fa46 Rename fma4 intrinsics to just fma since they are now used for both FMA4 and FMA3. Autoupgrade support coming in a separate commit.
llvm-svn: 157898
2012-06-03 07:26:46 +00:00
Stepan Dyatkovskiy
aae17ee48f Added unittests for IntegersSubset and IntegersSubsetMapping.
- Fixed IntegersSubsetGeneric copy/assignment behaviour. 
- Fixed IntegersSubsetGeneric::getSize/getSingleValue methods.
- Fixed IntegersSubsetGeneric::verify method.

Also IntegersSubset.h and IntegersSubsetMapping.h headers was fixed.

llvm-svn: 157887
2012-06-02 13:47:12 +00:00
Benjamin Kramer
bb30e1face Fix typos found by http://github.com/lyda/misspell-check
llvm-svn: 157885
2012-06-02 10:20:22 +00:00
Stepan Dyatkovskiy
4246540c22 PR1255: case ranges.
IntRange converted from struct to class. So main change everywhere is replacement of ".Low/High" with ".getLow/getHigh()"

llvm-svn: 157884
2012-06-02 09:42:43 +00:00
Benjamin Kramer
23f7f57c50 Add move semantics to APInt.
llvm-svn: 157883
2012-06-02 08:39:08 +00:00
Stepan Dyatkovskiy
1ed91c6a4b Additional change for 157881. Forget to fix another IntegerSubset constructor.
llvm-svn: 157882
2012-06-02 08:03:34 +00:00
Stepan Dyatkovskiy
8aeda74069 Small fix due to buildbot failures on mingw32. Fixed call of parent constructor for case when parent is template.
llvm-svn: 157881
2012-06-02 07:44:19 +00:00
Stepan Dyatkovskiy
f302516204 PR1255: case ranges.
IntegersSubsetGeneric, IntegersSubsetMapping: added IntTy template parameter, that allows use either APInt or IntItem. This change allows to write unittest for these classes.

llvm-svn: 157880
2012-06-02 07:26:00 +00:00
Jakob Stoklund Olesen
ee64ad3aba Remove the old register list functions from MCRegisterInfo.
These functions exposed the layout of the underlying data tables as
null-terminated uint16_t arrays.

Use the new MCSubRegIterator, MCSuperRegIterator, and MCRegAliasIterator
classes instead.

llvm-svn: 157855
2012-06-01 23:28:34 +00:00
Jakob Stoklund Olesen
be0b8939c0 Switch all register list clients to the new MC*Iterator interface.
No functional change intended.

Sorry for the churn. The iterator classes are supposed to help avoid
giant commits like this one in the future. The TableGen-produced
register lists are getting quite large, and it may be necessary to
change the table representation.

This makes it possible to do so without changing all clients (again).

llvm-svn: 157854
2012-06-01 23:28:30 +00:00
Manman Ren
2a5898a76b ARM: properly handle alignment for struct byval.
Factor out the expansion code into a function.
This change is to be enabled in clang.

rdar://9877866

llvm-svn: 157830
2012-06-01 19:33:18 +00:00
Benjamin Kramer
b9ec87406a Provide move semantics for (Small)BitVector.
CodeGen makes a lot of BitVector copies.

llvm-svn: 157826
2012-06-01 18:52:53 +00:00
Stepan Dyatkovskiy
13bc729662 PR1255: case ranges.
IntegersSubset devided into IntegersSubsetGeneric and into IntegersSubset itself. The first has no references to ConstantInt and works with IntItem only.
IntegersSubsetMapping also made generic. Here added second template parameter "IntegersSubsetTy" that allows to use on of two IntegersSubset types described below.

llvm-svn: 157815
2012-06-01 16:17:57 +00:00
Benjamin Kramer
5c03529e44 Remove noisy semicolons.
llvm-svn: 157814
2012-06-01 15:40:53 +00:00
Stepan Dyatkovskiy
4f89bda8fb PR1255: case ranges.
IntItem cleanup. IntItemBase, IntItemConstantIntImp and IntItem merged into IntItem. All arithmetic operators was propogated from APInt. Also added comparison operators <,>,<=,>=. Currently you will find set of macros that propogates operators from APInt to IntItem in the beginning of IntegerSubset. Note that THESE MACROS WILL REMOVED after all passes will case-ranges compatible. Also note that these macros much smaller pain that something like this:
if (V->getValue().ugt(AnotherV->getValue()) { ... }

These changes made IntItem full featured integer object. It allows to make IntegerSubset class generic (move out all ConstantInt references inside and add unit-tests) in next commits.

llvm-svn: 157810
2012-06-01 10:06:14 +00:00
Eric Christopher
2cff8a57a0 Add support for enum forward declarations.
Part of rdar://11570854

llvm-svn: 157786
2012-06-01 00:22:32 +00:00
Benjamin Kramer
2f57459340 IntrusiveRefCntPtr: Simplify operator= as suggested by Richard Smith.
This way the constructors do all the hard work. No intended functionality change.

llvm-svn: 157773
2012-05-31 22:25:25 +00:00
Manman Ren
82e2c9debf X86: replace SUB with CMP if possible
This patch will optimize the following
        movq    %rdi, %rax
        subq    %rsi, %rax
        cmovsq  %rsi, %rdi
        movq    %rdi, %rax
to
        cmpq    %rsi, %rdi
        cmovsq  %rsi, %rdi
        movq    %rdi, %rax

Perform this optimization if the actual result of SUB is not used.

rdar: 11540023
llvm-svn: 157755
2012-05-31 17:20:29 +00:00
Jakob Stoklund Olesen
dc9a46b651 Add a PrintRegUnit helper similar to PrintReg.
Reg-units are named after their root registers, and most units have a
single root, so they simply print as 'AL', 'XMM0', etc. The rare dual
root reg-units print as FPSCR~FPSCR_NZCV, FP0~ST7, ...

The printing piggybacks on the existing register name tables, so no
extra const data space is required.

llvm-svn: 157754
2012-05-31 17:18:29 +00:00
Jakob Stoklund Olesen
e79316648d Emit register unit root tables.
Each register unit has one or two root registers. The full set of
registers containing a given register unit can be computed as the union
of the root registers and their super-registers.

Provide an MCRegUnitRootIterator class to enumerate the roots.

llvm-svn: 157753
2012-05-31 17:18:26 +00:00
Craig Topper
c6bd90e646 Add intrinsic for pclmulqdq instruction.
llvm-svn: 157731
2012-05-31 04:37:40 +00:00
Benjamin Kramer
3578c169bf Teach SCEV's icmp simplification logic that a-b == 0 is equivalent to a == b.
This also required making recursive simplifications until
nothing changes or a hard limit (currently 3) is hit.

With the simplification in place indvars can canonicalize
loops of the form
for (unsigned i = 0; i < a-b; ++i)
into
for (unsigned i = 0; i != a-b; ++i)
which used to fail because SCEV created a weird umax expr
for the backedge taken count.

llvm-svn: 157701
2012-05-30 18:32:23 +00:00
Jakob Stoklund Olesen
67b7a47013 Add MCRegisterInfo::RegListIterator.
Also add subclasses MCSubRegIterator, MCSuperRegIterator, and
MCRegAliasIterator.

These iterators provide an abstract interface to the MCRegisterInfo
register lists so the internal representation can be changed without
changing all clients.

llvm-svn: 157695
2012-05-30 16:36:28 +00:00
Benjamin Kramer
df046ac8f9 Mark insertq/extrq intrinsic readnone.
llvm-svn: 157688
2012-05-30 13:44:25 +00:00
Bob Wilson
387d610b8d Add an insertPass API to TargetPassConfig. <rdar://problem/11498613>
Besides adding the new insertPass function, this patch uses it to
enhance the existing -print-machineinstrs so that the MachineInstrs
after a specific pass can be printed.

Patch by Bin Zeng!

llvm-svn: 157655
2012-05-30 00:17:12 +00:00
Jakob Stoklund Olesen
73535ebd9d Make DiffListIterator public to unbreak the gcc buildbots.
Apparently, a friend can't derive from a private class according to gcc.

llvm-svn: 157654
2012-05-30 00:05:03 +00:00
Jakob Stoklund Olesen
c83fa0a0e4 Use MCRegUnitIterator to compute regsOverlap().
The register unit lists are typically much shorter than the register
overlap lists, and the backing table for register units has better cache
locality because it is smaller.

This makes llc about 0.5% faster. The regsOverlap() function isn't that hot.

llvm-svn: 157651
2012-05-29 23:40:02 +00:00
Jakob Stoklund Olesen
16603c4371 Emit register unit lists for each register.
Register units are already used internally in TableGen to compute
register pressure sets and overlapping registers. This patch makes them
available to the code generators.

The register unit lists are differentially encoded so they can be reused
for many related registers. This keeps the total size of the lists below
200 bytes for most targets. ARM has the largest table at 560 bytes.

Add an MCRegUnitIterator for traversing the register unit lists. It
provides an abstract interface so the representation can be changed in
the future without changing all clients.

llvm-svn: 157650
2012-05-29 23:40:00 +00:00
Douglas Gregor
bbf03d028e DenseMap's move assignment operator needs to return *this
llvm-svn: 157644
2012-05-29 20:33:09 +00:00
Benjamin Kramer
0c823ae0ed Add intrinsics, code gen, assembler and disassembler support for the SSE4a extrq and insertq instructions.
This required light surgery on the assembler and disassembler
because the instructions use an uncommon encoding. They are
the only two instructions in x86 that use register operands
and two immediates.

llvm-svn: 157634
2012-05-29 19:05:25 +00:00
Stepan Dyatkovskiy
23fa9fd605 ConstantRangesSet renamed to IntegersSubset. CRSBuilder renamed to IntegersSubsetMapping.
llvm-svn: 157612
2012-05-29 12:26:47 +00:00
Peter Collingbourne
b5121f5f06 Add llvm.fabs intrinsic.
llvm-svn: 157594
2012-05-28 21:48:37 +00:00
Meador Inge
af8619b4de PR12696: Attribute bits above 1<<30 are not encoded in bitcode
Attribute bits above 1<<30 are now encoded correctly.  Additionally,
the encoding/decoding functionality has been hoisted to helper functions
in Attributes.h in an effort to help the encoding/decoding to stay in
sync with the Attribute bitcode definitions.

llvm-svn: 157581
2012-05-28 15:45:43 +00:00
Stepan Dyatkovskiy
d534e24968 PR1255: Case Ranges
Implemented IntItem - the wrapper around APInt. Why not to use APInt item directly right now?
1. It will very difficult to implement case ranges as series of small patches. We got several large and heavy patches. Each patch will about 90-120 kb. If you replace ConstantInt with APInt in SwitchInst you will need to changes at the same time all Readers,Writers and absolutely all passes that uses SwitchInst.
2. We can implement APInt pool inside and save memory space. E.g. we use several switches that works with 256 bit items (switch on signatures, or strings). We can avoid value duplicates in this case.
3. IntItem can be easyly easily replaced with APInt.
4. Currenly we can interpret IntItem both as ConstantInt and as APInt. It allows to provide SwitchInst methods that works with ConstantInt for non-updated passes.

Why I need it right now? Currently I need to update SimplifyCFG pass (EqualityComparisons). I need to work with APInts directly a lot, so peaces of code
ConstantInt *V = ...;
if (V->getValue().ugt(AnotherV->getValue()) {
  ...
}
will look awful. Much more better this way:
IntItem V = ConstantIntVal->getValue();
if (AnotherV < V) {
}

Of course any reviews are welcome.

P.S.: I'm also going to rename ConstantRangesSet to IntegersSubset, and CRSBuilder to IntegersSubsetMapping (allows to map individual subsets of integers to the BasicBlocks).
Since in future these classes will founded on APInt, it will possible to use them in more generic ways.

llvm-svn: 157576
2012-05-28 12:39:09 +00:00
Stepan Dyatkovskiy
3f34a2e3a1 SwitchInst: Due to bad readability case iterators definition was moved to the end of SwitchInst.
llvm-svn: 157575
2012-05-28 10:11:27 +00:00
Chris Lattner
afebb9fced switch AttrListPtr::get to take an ArrayRef, simplifying a lot of clients.
llvm-svn: 157556
2012-05-28 01:47:44 +00:00
Chris Lattner
6cdc78cdf4 add some helper methods to make the type more uniform.
llvm-svn: 157554
2012-05-28 01:29:59 +00:00
Chris Lattner
b051169784 rdar://11542750 - llvm.trap should be marked no return.
llvm-svn: 157551
2012-05-27 23:20:41 +00:00
Benjamin Kramer
fb377a947e DenseMap: Use an early exit when there is nothing to do in DestroyAll().
llvm-svn: 157550
2012-05-27 22:53:10 +00:00
Benjamin Kramer
5144cce317 IntrusiveRefCntPtr: Use the same pattern as the other operator= overloads when using rvalue refs.
llvm-svn: 157546
2012-05-27 20:46:04 +00:00
Chris Lattner
66230fbd9c generalize this to allow any argument.
llvm-svn: 157542
2012-05-27 19:17:16 +00:00
Chris Lattner
073117cd52 move some code around so that Verifier.cpp can get access to the intrinsic info table.
llvm-svn: 157540
2012-05-27 18:28:35 +00:00
Benjamin Kramer
895da0f01a DenseMap: Provide a move ctor and move semantics for operator[] and FindAndConstruct.
The only missing part is insert(), which uses a pair of parameters and I haven't
figured out how to convert it to rvalue references. It's now possible to use a
DenseMap with std::unique_ptr values :)

llvm-svn: 157539
2012-05-27 17:38:30 +00:00
Benjamin Kramer
ceee3e117e DenseMap: Factor destruction into a common helper method.
llvm-svn: 157538
2012-05-27 17:38:18 +00:00
Benjamin Kramer
c529ae7728 Move-enable IntrusiveRefCntPtr.
These tend to be copied around a lot, moving it instead saves a ton of memory
accesses.

llvm-svn: 157535
2012-05-27 16:22:08 +00:00
Benjamin Kramer
0284253cb3 Add support for branch weight metadata to MDBuilder and use it in various places.
llvm-svn: 157515
2012-05-26 13:59:43 +00:00
Justin Holewinski
77c4679dae Change interface for TargetLowering::LowerCallTo and TargetLowering::LowerCall
to pass around a struct instead of a large set of individual values.  This
cleans up the interface and allows more information to be added to the struct
for future targets without requiring changes to each and every target.

NV_CONTRIB

llvm-svn: 157479
2012-05-25 16:35:28 +00:00
Jakob Stoklund Olesen
87e444d0a5 Compress MCRegisterInfo register name tables.
Store (debugging) register names as offsets into a string table instead
of as char pointers.

llvm-svn: 157449
2012-05-25 00:21:41 +00:00
Eli Friedman
d89582030a Simplify code for calling a function where CanLowerReturn fails, fixing a small bug in the process.
llvm-svn: 157446
2012-05-25 00:09:29 +00:00
Jakob Stoklund Olesen
4d1e62df51 Don't put TGParser scratch results in the output.
Only fully expanded Records should go into RecordKeeper.

llvm-svn: 157431
2012-05-24 22:17:36 +00:00
Andrew Trick
3152745a8f misched: Added ScoreboardHazardRecognizer.
The Hazard checker implements in-order contraints, or interlocked
resources. Ready instructions with hazards do not enter the available
queue and are not visible to other heuristics.

The major code change is the addition of SchedBoundary to encapsulate
the state at the top or bottom of the schedule, including both a
pending and available queue.

The scheduler now counts cycles in sync with the hazard checker. These
are minimum cycle counts based on known hazards.

Targets with no itinerary (x86_64) currently remain at cycle 0. To fix
this, we need to provide some maximum issue width for all targets. We
also need to add the concept of expected latency vs. minimum latency.

llvm-svn: 157427
2012-05-24 22:11:09 +00:00
Justin Holewinski
ee3de35dee Remove the PTX back-end and all of its artifacts (triple, etc.)
This back-end was deprecated in favor of the NVPTX back-end.

NV_CONTRIB

llvm-svn: 157417
2012-05-24 21:38:21 +00:00
Owen Anderson
03e1ab9c5b Teach tblgen's set theory "sequence" operator to support an optional stride operand.
llvm-svn: 157416
2012-05-24 21:37:08 +00:00
Nuno Lopes
68aaef6b21 revert r156383: removal of TYPE_CODE_FUNCTION_OLD
Apparently LLVM only stopped emitting this after LLVM 3.0

llvm-svn: 157325
2012-05-23 15:19:39 +00:00
Eric Christopher
809a39bd54 Add support for C++11 enum classes in llvm.
Part of rdar://11496790

llvm-svn: 157303
2012-05-23 00:09:20 +00:00
Nuno Lopes
114b8eaa9c add a new pass to instrument loads and stores for run-time bounds checking
move EmitGEPOffset from InstCombine to Transforms/Utils/Local.h

(a draft of this) patch reviewed by Andrew, thanks.

llvm-svn: 157261
2012-05-22 17:19:09 +00:00
Nuno Lopes
944814b41a revert my previous patches that introduced an additional parameter to the objectsize intrinsic.
After a lot of discussion, we realized it's not the best option for run-time bounds checking

llvm-svn: 157255
2012-05-22 15:25:31 +00:00
Nuno Lopes
328f9efec2 fix the quotient returned by sdivrem() for the case when LHS is negative and RHS is positive
based on a patch by Preston Briggs, with some modifications

llvm-svn: 157231
2012-05-22 01:09:48 +00:00
Pete Cooper
869fc085f9 Added address space qualifier to intrinsic PointerType arguments.
llvm-svn: 157218
2012-05-21 23:21:28 +00:00
Stepan Dyatkovskiy
98d41ce6af PR1255 (case ranges: work with ConstantRangesSet instead of ConstantInt) related changes for Execution and Verifier.
llvm-svn: 157183
2012-05-21 10:44:40 +00:00
Jakob Stoklund Olesen
c058c41261 Give a small negative bias to giant edge bundles.
This helps compile time when the greedy register allocator splits live
ranges in giant functions. Without the bias, we would try to grow
regions through the giant edge bundles, usually to find out that the
region became too big and expensive.

If a live range has many uses in blocks near the giant bundle, the small
negative bias doesn't make a big difference, and we still consider
regions including the giant edge bundle.

Giant edge bundles are usually connected to landing pads or indirect
branches.

llvm-svn: 157174
2012-05-21 03:11:23 +00:00
Jakob Stoklund Olesen
178f5ce549 Add a LiveRangeQuery class.
This class is meant to be the primary interface for examining a live
range in the vicinity on a given instruction. It avoids all the messy
dealings with iterators and early clobbers.

This is a more abstract interface to live ranges, hiding the
implementation as a vector of segments.

llvm-svn: 157141
2012-05-20 02:44:30 +00:00
Benjamin Kramer
386e9a3f75 Disambiguate call to operator==.
clang++ and msvc happily had no problem with it but g++ refuses to compile.

llvm-svn: 157126
2012-05-19 19:32:11 +00:00
Benjamin Kramer
c4cb78c727 ValueMap: Use DenseMap's find_as mechanism to reduce use list churn.
Otherwise just looking up a value in the map requires creating a VH, adding it to the use lists and destroying it again.

llvm-svn: 157124
2012-05-19 19:15:32 +00:00
Benjamin Kramer
e11ffa2475 Move CallbackVHs dtor inline, it can be devirtualized in many cases. Move the other virtual methods out of line as they are only called from within Value.cpp anyway.
llvm-svn: 157123
2012-05-19 19:15:25 +00:00
Benjamin Kramer
16ca97c2f1 Provide move semantics for TinyPtrVector and for DenseMap's rehash function.
This makes DenseMap<..., TinyPtrVector<...>> as cheap as it always should've been!

llvm-svn: 157113
2012-05-19 13:28:54 +00:00
Jakob Stoklund Olesen
6a5bbcc25c Allow LiveRangeEdit to be created with a NULL parent.
The dead code elimination with callbacks is still useful.

llvm-svn: 157100
2012-05-19 05:25:46 +00:00
Eric Christopher
74be9f0f18 Actually support DW_TAG_rvalue_reference_type that we were trying
to generate out of the front end.

rdar://11479676

llvm-svn: 157094
2012-05-19 01:36:37 +00:00
Andrew Trick
c3765e6af0 SCEV: Add MarkPendingLoopPredicates to avoid recursive isImpliedCond.
getUDivExpr attempts to simplify by checking for overflow.
isLoopEntryGuardedByCond then evaluates the loop predicate which
may lead to the same getUDivExpr causing endless recursion.

Fixes PR12868: clang 3.2 segmentation fault.

llvm-svn: 157092
2012-05-19 00:48:25 +00:00
Jakob Stoklund Olesen
ffe87e7e00 Modernize naming convention for class members.
No functional change.

llvm-svn: 157079
2012-05-18 22:10:15 +00:00
Jim Grosbach
343a996ca5 Refactor data-in-code annotations.
Use a dedicated MachO load command to annotate data-in-code regions.
This is the same format the linker produces for final executable images,
allowing consistency of representation and use of introspection tools
for both object and executable files.

Data-in-code regions are annotated via ".data_region"/".end_data_region"
directive pairs, with an optional region type.

data_region_directive := ".data_region" { region_type }
region_type := "jt8" | "jt16" | "jt32" | "jta32"
end_data_region_directive := ".end_data_region"

The previous handling of ARM-style "$d.*" labels was broken and has
been removed. Specifically, it didn't handle ARM vs. Thumb mode when
marking the end of the section.

rdar://11459456

llvm-svn: 157062
2012-05-18 19:12:01 +00:00
Nick Kledzik
d21d4e5653 fix warnings when compiling with -Wshadow
llvm-svn: 157061
2012-05-18 18:39:06 +00:00
Evandro Menezes
740dea501d [Hexagon] Clean up Hexagon ELF definition.
llvm-svn: 156996
2012-05-17 16:46:46 +00:00
Eric Christopher
4c15973dff Grammar.
llvm-svn: 156955
2012-05-16 22:08:58 +00:00
Duncan Sands
cd6958362b I noticed that named metadata doesn't provide a direct way of getting at the
named metadata list, unlike all the other global objects (global variables,
functions, aliases), so add that for consistency.

llvm-svn: 156915
2012-05-16 12:25:43 +00:00
Daniel Dunbar
35f13a8fee [Support] Add a version of sys::fs::equivalent() that treats errors as false.
llvm-svn: 156864
2012-05-15 22:07:14 +00:00
Jim Grosbach
8f241263d7 TableGen'erate mapping physical registers to encoding values.
Many targets always use the same bitwise encoding value for physical
registers in all (or most) instructions. Add this mapping to the
.td files and TableGen'erate the information and expose an accessor
in MCRegisterInfo.

patch by Tom Stellard.

llvm-svn: 156829
2012-05-15 17:35:57 +00:00
Jim Grosbach
2e62e2f664 Allow MCCodeEmitter access to the target MCRegisterInfo.
Add the MCRegisterInfo to the factories and constructors.

Patch by Tom Stellard <Tom.Stellard@amd.com>.

llvm-svn: 156828
2012-05-15 17:35:52 +00:00
Stepan Dyatkovskiy
6d83dd3cf8 Fixed one small stupid, but critical bug.
llvm-svn: 156810
2012-05-15 09:21:39 +00:00
Michael J. Spencer
5a53bd3002 [Support/COFF] Make the order of members in symbol match the standard.
llvm-svn: 156785
2012-05-14 22:43:21 +00:00
Chad Rosier
c3a90c47b9 Move the capture analysis from MemoryDependencyAnalysis to a more general place
so that it can be reused in MemCpyOptimizer.  This analysis is needed to remove
an unnecessary memcpy when returning a struct into a local variable.
rdar://11341081
PR12686

llvm-svn: 156776
2012-05-14 20:35:04 +00:00
Dan Gohman
cc1f60a86c Rename @llvm.debugger to @llvm.debugtrap.
llvm-svn: 156774
2012-05-14 18:58:10 +00:00
Andrew Trick
c0a8fc3638 Remove a stale forward declaration.
llvm-svn: 156770
2012-05-14 18:03:19 +00:00
Jakob Stoklund Olesen
184e921187 Remove the expensive BitVector::operator~().
Returning a temporary BitVector is very expensive. If you must, create
the temporary explicitly: Use BitVector(A).flip() instead of ~A.

llvm-svn: 156768
2012-05-14 15:46:27 +00:00
Jakob Stoklund Olesen
f8648a8b0d Remove BitVector binops.
These operators were crazy slow, calling malloc to return a temporary
result. At the same time, they look very innocent when used in code.

If you need temporary BitVectors to compute your thing, create them
explicitly, and use the inplace logical operators. This makes the high
cost explicit in the code.

llvm-svn: 156767
2012-05-14 15:37:25 +00:00
Jakob Stoklund Olesen
8783cdf045 Add BitVector::anyCommon().
The existing operation (A & B).any() is very slow.

llvm-svn: 156760
2012-05-14 15:01:19 +00:00
Stepan Dyatkovskiy
f050e5bcbf SwitchInst cosmetics: renamed "Hash" method to "hash"
llvm-svn: 156757
2012-05-14 08:26:31 +00:00
Bill Wendling
aad3af5f7c Use ArrayRef instead of an explicit vector type.
llvm-svn: 156755
2012-05-14 07:53:40 +00:00
Stepan Dyatkovskiy
fa0cf8dc2e Recommited r156374 with critical fixes in BitcodeReader/Writer:
Ordinary patch for PR1255.
Added new case-ranges orientated methods for adding/removing cases in SwitchInst. After this patch cases will internally representated as ConstantArray-s instead of ConstantInt, externally cases wrapped within the ConstantRangesSet object.
Old methods of SwitchInst are also works well, but marked as deprecated. So on this stage we have no side effects except that I added support for case ranges in BitcodeReader/Writer, of course test for Bitcode is also added. Old "switch" format is also supported.

llvm-svn: 156704
2012-05-12 10:48:17 +00:00
Jay Foad
65d25fa204 Teach Function::hasAddressTaken that BlockAddress doesn't really take
the address of a function.

llvm-svn: 156703
2012-05-12 08:30:16 +00:00
Michael J. Spencer
ca721ac5bf Add doxygen comments.
llvm-svn: 156665
2012-05-11 23:34:39 +00:00
Michael J. Spencer
6161587c9f [Support/StringRef] Add find_last_not_of and {r,l,}trim.
llvm-svn: 156652
2012-05-11 22:08:50 +00:00
Sirish Pande
4590b341e2 Hexagon V5 intrinsics support.
llvm-svn: 156631
2012-05-11 19:39:13 +00:00
Stepan Dyatkovskiy
a1652c65c7 PR1255: ConstantRangesSet and CRSBuilder classes moved from include/llvm to include/llvm/Support.
llvm-svn: 156613
2012-05-11 10:34:23 +00:00
Jim Grosbach
f588df5936 Tidy up. Trailing whitespace.
llvm-svn: 156601
2012-05-11 01:39:13 +00:00
Dan Gohman
ed475ad173 Define a new intrinsic, @llvm.debugger. It will be similar to __builtin_trap(),
but it generates int3 on x86 instead of ud2.

llvm-svn: 156593
2012-05-11 00:19:32 +00:00
Eric Christopher
c2a56d46fc Allow unique_file to take a mode for file permissions, but default
to user only read/write.

Part of rdar://11325849

llvm-svn: 156591
2012-05-11 00:07:44 +00:00
Dan Gohman
8b1a3cec89 Teach DeadStoreElimination to eliminate exit-block stores with phi addresses.
llvm-svn: 156558
2012-05-10 18:57:38 +00:00
Chad Rosier
219f1f7ac8 Add Triple::getiOSVersion.
This new function provides a way to get the iOS version number from ios triples.
Part of rdar://11409204

llvm-svn: 156483
2012-05-09 17:23:48 +00:00
Hans Wennborg
879332e389 Introduce llvm-c function LLVMPrintModuleToFile.
This lets you save the textual representation of the LLVM IR to a file.
Before this patch it could only be printed to STDERR from llvm-c.

Patch by Carlo Kok!

llvm-svn: 156479
2012-05-09 16:54:17 +00:00
Nuno Lopes
e8880a9916 change the objectsize intrinsic signature: add a 3rd parameter to denote the maximum runtime performance penalty that the user is willing to accept.
This commit only adds the parameter. Code taking advantage of it will follow.

llvm-svn: 156473
2012-05-09 15:52:43 +00:00
Bill Wendling
4e03ce2c3c Supply a C interface to the "LinkModules" method.
Patch by Andrew Wilkins!

llvm-svn: 156469
2012-05-09 08:55:40 +00:00
Kevin Enderby
97a05ec660 Fix it so llvm-objdump -arch does accept x86 and x86-64 as valid arch names.
PR12731.  Patch by Meador Inge!

llvm-svn: 156444
2012-05-08 23:38:45 +00:00
Eric Christopher
255767a1d7 Remove excess semi-colons to quiet warnings.
llvm-svn: 156416
2012-05-08 20:45:04 +00:00
Eric Christopher
57d125a5a0 Update comment.
llvm-svn: 156404
2012-05-08 18:55:57 +00:00
Nuno Lopes
7b6d75a852 remove TYPE_CODE_FUNCTION_OLD type code. it is no longer in use and it was marked for removal in 3.0
llvm-svn: 156383
2012-05-08 16:16:20 +00:00
Stepan Dyatkovskiy
b150cd5ced Rejected r156374: Ordinary PR1255 patch. Due to clang-x86_64-debian-fnt buildbot failure.
llvm-svn: 156377
2012-05-08 08:33:21 +00:00
Craig Topper
77b1a4cee5 Remove 256-bit AVX non-temporal store intrinsics. Similar was previously done for 128-bit.
llvm-svn: 156375
2012-05-08 06:58:15 +00:00
Stepan Dyatkovskiy
33fd2a5bf4 Ordinary patch for PR1255.
Added new case-ranges orientated methods for adding/removing cases in SwitchInst. After this patch cases will internally representated as ConstantArray-s instead of ConstantInt, externally cases wrapped within the ConstantRangesSet object.
Old methods of SwitchInst are also works well, but marked as deprecated. So on this stage we have no side effects except that I added support for case ranges in BitcodeReader/Writer, of course test for Bitcode is also added. Old "switch" format is also supported.

llvm-svn: 156374
2012-05-08 06:36:08 +00:00
Jakob Stoklund Olesen
cc0cf22b98 Add an MF argument to TRI::getPointerRegClass() and TII::getRegClass().
The getPointerRegClass() hook can return register classes that depend on
the calling convention of the current function (ptr_rc_tailcall).

So far, we have been able to infer the calling convention from the
subtarget alone, but as we add support for multiple calling conventions
per target, that no longer works.

Patch by Yiannis Tsiouris!

llvm-svn: 156328
2012-05-07 22:10:26 +00:00
Jordy Rose
7d6fc98ccf Constify (trivially) ImmutableSet::iterator::getVisitState().
This was probably intended all along.

llvm-svn: 156318
2012-05-07 19:24:40 +00:00
Jakob Stoklund Olesen
d92280013d Add TRI::getCommonSuperRegClass().
This function is a generalization of getMatchingSuperRegClass() to the
symmetric case where both sides are using a sub-register index. It will
find a super-register class and sub-register indexes that make this
diagram commute:

                                   PreA
                       SuperRC  ---------->  RCA

                          |                   |
                          |                   |
                     PreB |                   | SubA
                          |                   |
                          |                   |
                          V                   V

                         RCB    ----------> SubRC
                                   SubB

This can be used to coalesce copies like:

  %vreg1:sub16 = COPY %vreg2:sub16; GR64:%vreg1, GR32: %vreg2

llvm-svn: 156317
2012-05-07 19:14:58 +00:00
John McCall
98163678d4 Fix trivial typo in llvm_move.
llvm-svn: 156288
2012-05-07 06:00:23 +00:00
Craig Topper
c6d0bc2afc Add SSE4A MOVNTSS/MOVNTSD instructions.
llvm-svn: 156281
2012-05-07 05:36:19 +00:00
Jim Grosbach
970a443ab5 Tidy up. Whitespace.
llvm-svn: 156276
2012-05-07 02:25:53 +00:00
Chris Lattner
cf8284517f reapply my patch, with a fix for an off-by-one error. Turned out to be a lot
of work for a drive-by fix :)

llvm-svn: 156246
2012-05-05 22:17:32 +00:00
Chris Lattner
206bf447c0 revert my patches, which are causing problems.
llvm-svn: 156245
2012-05-05 22:11:04 +00:00
Chris Lattner
372a67f8d9 add missing header <shame>
llvm-svn: 156244
2012-05-05 22:04:11 +00:00
Jim Grosbach
f7461026c2 Nuke a few dead remnants of the CBE.
llvm-svn: 156241
2012-05-05 17:45:12 +00:00
Daniel Dunbar
457eab2ad7 [Support] Add sys::Process::GetRandomNumber().
- Primitive API, but we rarely have need for random numbers.

llvm-svn: 156237
2012-05-05 16:36:20 +00:00
Daniel Dunbar
c69f7c9234 [build] Add build check for ::arc4random().
llvm-svn: 156236
2012-05-05 16:36:16 +00:00
Benjamin Kramer
7a9528b540 Add a new target hook "predictableSelectIsExpensive".
This will be used to determine whether it's profitable to turn a select into a
branch when the branch is likely to be predicted.

Currently enabled for everything but Atom on X86 and Cortex-A9 devices on ARM.

I'm not entirely happy with the name of this flag, suggestions welcome ;)

llvm-svn: 156233
2012-05-05 12:49:14 +00:00