Commit Graph

6306 Commits

Author SHA1 Message Date
Dan Gohman
861923ed06 From Chris' review: change MemOperandSDNode's constructor to pass its
argument by reference, rather than by value. 

llvm-svn: 46960
2008-02-11 18:56:50 +00:00
Ted Kremenek
f06b7ac532 Added "Profile" method to APFloat for use with FoldingSet.
Added member template "Add" to FoldingSetNodeID that allows "adding" arbitrary
objects to a profile via dispatch to FoldingSetTrait<T>::Profile().

Removed FoldingSetNodeID::AddAPFloat and FoldingSetNodeID::APInt, as their
functionality is now replaced using the above mentioned member template.

llvm-svn: 46957
2008-02-11 17:24:50 +00:00
Duncan Sands
b9bf0dcb7c Add arbitrary integer support to getRegisterType and
getNumRegisters.  This is needed for calling functions
with apint parameters or return values.

llvm-svn: 46956
2008-02-11 11:09:23 +00:00
Wojciech Matyjewicz
76c2b22ee4 Fix PR1798 - an error in the evaluation of SCEVAddRecExpr at an
arbitrary iteration.

The patch:
1) changes SCEVSDivExpr into SCEVUDivExpr,
2) replaces PartialFact() function with BinomialCoefficient(); the 
computations (essentially, the division) in BinomialCoefficient() are 
performed with the apprioprate bitwidth necessary to avoid overflow; 
unsigned division is used instead of the signed one.

Computations in BinomialCoefficient() require support from the code 
generator for APInts. Currently, we use a hack rounding up the 
neccessary bitwidth to the nearest power of 2. The hack is easy to turn 
off in future.

One remaining issue: we assume the divisor of the binomial coefficient 
formula can be computed accurately using 16 bits. It means we can handle 
AddRecs of length up to 9. In future, we should use APInts to evaluate 
the divisor.

Thanks to Nicholas for cooperation!

llvm-svn: 46955
2008-02-11 11:03:14 +00:00
Duncan Sands
204c89cafa Add a isBigEndian method to complement isLittleEndian.
llvm-svn: 46954
2008-02-11 10:37:04 +00:00
Dan Gohman
cabaec582f Rename MRegisterInfo to TargetRegisterInfo.
llvm-svn: 46930
2008-02-10 18:45:23 +00:00
Wojciech Matyjewicz
2e45476439 Add a check if the initial value of the induction variable is 0 (the method comment says it should be).
llvm-svn: 46907
2008-02-09 18:06:58 +00:00
Dan Gohman
672d341167 Change ConstantSDNode to store an APInt instead of a uint64_t, and
begin adding some methods to use it this way.

llvm-svn: 46899
2008-02-08 22:59:30 +00:00
Dan Gohman
3d296e049c Add an isSignedIntN, like isIntN but for signed integer values instead of
unsigned.

llvm-svn: 46894
2008-02-08 21:58:15 +00:00
Evan Cheng
90f03a0b88 It's not always safe to fold movsd into xorpd, etc. Check the alignment of the load address first to make sure it's 16 byte aligned.
llvm-svn: 46893
2008-02-08 21:20:40 +00:00
Ted Kremenek
557bbab669 Constified operator<< in APSInt.
llvm-svn: 46882
2008-02-08 07:14:19 +00:00
Dan Gohman
1f255f2db7 Avoid needlessly casting away const qualifiers.
llvm-svn: 46876
2008-02-08 03:26:46 +00:00
Dan Gohman
eb7c8e4f6b Follow Chris' suggestion; change the PseudoSourceValue accessors
to return pointers instead of references, since this is always what
is needed.

llvm-svn: 46857
2008-02-07 18:41:25 +00:00
Dan Gohman
011075cfdb Add support to FoldingSet for hashing APInt objects.
llvm-svn: 46833
2008-02-06 23:09:15 +00:00
Dan Gohman
f00842e086 Re-apply the memory operand changes, with a fix for the static
initializer problem, a minor tweak to the way the
DAGISelEmitter finds load/store nodes, and a renaming of the
new PseudoSourceValue objects.

llvm-svn: 46827
2008-02-06 22:27:42 +00:00
Anton Korobeynikov
b1576212e7 Don't dereference an invalid pointer if string is empty.
llvm-svn: 46781
2008-02-05 23:34:40 +00:00
Dale Johannesen
e8fa3130f8 Implement sseregparm.
llvm-svn: 46764
2008-02-05 20:46:33 +00:00
Ted Kremenek
3277f38b25 Added FoldingSet profiling support to ImmutableSet.
llvm-svn: 46757
2008-02-05 18:50:25 +00:00
Ted Kremenek
571ad3913c Changed profiling method for ImmutableMap to once again just use its
unique ImutAVLTree* for profiling.

Modified ImutAVLTree:
 (1) changed ComputeHash() to ComputeDigest() and
 (2) changed Profile() to use the computed digest and
 (3) modified insertion of IMutAVLTree into the FoldingSet owned by
     the ImutAVLTreeFactory object to use profiling instead of computing
     a direct hash.  This fixes a bug where our abuse of the FoldingSet would
     not work when the FoldingSet was resized.

llvm-svn: 46753
2008-02-05 17:30:43 +00:00
Dale Johannesen
bf7691e2e8 Missed one.
llvm-svn: 46733
2008-02-05 01:12:10 +00:00
Evan Cheng
ca7db1ca34 Unbreak teh build.
llvm-svn: 46729
2008-02-05 00:25:13 +00:00
Dale Johannesen
9b7f4064e3 Do not unconditionally redefine vec_ext_v16qi and
vec_ext_v4si builtins.  This is a hack; they should
be defined here, then resolved in the X86 BE.
However there is enough other stuff missing in the
X86 BE for SSE41 that this will do for now.

llvm-svn: 46727
2008-02-04 23:27:29 +00:00
Evan Cheng
352062f5f7 Typo.
llvm-svn: 46725
2008-02-04 23:10:38 +00:00
Ted Kremenek
f2e96e37f4 Modified 'Profile' method of ImmutableMap to use the 'Profile' method
of the contained ImutAVLTree root.

llvm-svn: 46719
2008-02-04 21:17:02 +00:00
Ted Kremenek
9679f99479 Modified node creation of ImutAVLTree to do a hash lookup for an existing
node in the FoldingSet of nodes held by the Factory object.  If we
we find a node with a matching hash, we do a full structural comparison.
Nodes are also now inserted into the FoldingSet only when we mark them
Immutable, as their children can change during intermediate-rebalancing.

The 'Profile' method for ImutAVLTree is no longer used when looking up
existing ImutAVLTrees with a given set of contents; instead the Profile method
is used by other clients that wish to insert such a tree into a folding set.
This means that we are not using FoldingSet in ImutAVLTreeFactory in the way
it was intended, but instead are using it as an opaque hashtable.

llvm-svn: 46717
2008-02-04 21:15:24 +00:00
Ted Kremenek
6dab41ed32 Added "bucket_iterators" to FoldingSet. Bucket iterators allow iteration
over all the nodes in a particular bucket.

llvm-svn: 46716
2008-02-04 21:11:17 +00:00
Ted Kremenek
62853a3478 constified operator~().
llvm-svn: 46707
2008-02-04 16:57:26 +00:00
Nate Begeman
ead8dfeef2 SSE 4.1 Intrinsics and detection
llvm-svn: 46681
2008-02-03 07:18:54 +00:00
Chris Lattner
6ea106139d Change the 'global modification' APIs in SelectionDAG to take a new
DAGUpdateListener object pointer instead of just returning a vector 
of deleted nodes.  This makes the interfaces more efficient (no more
allocating a vector [at least a malloc], filling it in, then walking
it) and more clean.  This also allows the client to be notified of
nodes that are *changed* but not deleted.

llvm-svn: 46677
2008-02-03 06:49:24 +00:00
Evan Cheng
c57ec111f2 SDIsel processes llvm.dbg.declare by recording the variable debug information descriptor and its corresponding stack frame index in MachineModuleInfo. This only works if the local variable is "homed" in the stack frame. It does not work for byval parameter, etc.
Added ISD::DECLARE node type to represent llvm.dbg.declare intrinsic. Now the intrinsic calls are lowered into a SDNode and lives on through out the codegen passes.
For now, since all the debugging information recording is done at isel time, when a ISD::DECLARE node is selected, it has the side effect of also recording the variable. This is a short term solution that should be fixed in time.

llvm-svn: 46659
2008-02-02 04:07:54 +00:00
Gordon Henriksen
b75d9e974d Fixing a bug creating floating point constants of type other
than double through the C bindings. Thanks to Tomas Lindquist
Olsen for reporting it.

llvm-svn: 46656
2008-02-02 01:07:50 +00:00
Evan Cheng
9ff6b89bd9 Frame index can be negative.
llvm-svn: 46655
2008-02-02 00:17:00 +00:00
Evan Cheng
d6222fc11d Remove the nasty LABEL hack with a much less evil one. Now llvm.dbg.func.start implies a stoppoint is set. SelectionDAGISel records a new source line but does not create a ISD::LABEL node for this special stoppoint. Asm printer will magically print this label. This ensures nothing is emitted before.
llvm-svn: 46635
2008-02-01 09:10:45 +00:00
Evan Cheng
abe3a9ad27 Rename RecordLabel to RecordSourceLine because that's what it is doing.
llvm-svn: 46628
2008-02-01 02:05:57 +00:00
Anton Korobeynikov
1cc154d068 Add convenient std::string helpers to StringMap. Patch by Mikhail Glushenkov!
llvm-svn: 46625
2008-01-31 23:02:33 +00:00
Evan Cheng
2a533e6894 Revert 46556 and 46585. Dan please fix the PseudoSourceValue problem and re-commit.
llvm-svn: 46623
2008-01-31 21:00:00 +00:00
Anton Korobeynikov
91b7d99e8a Reapply the patch with fix.
llvm-svn: 46615
2008-01-31 17:43:25 +00:00
Chris Lattner
87c2925656 revert anton's recent stringmap patch, which breaks clang.
llvm-svn: 46614
2008-01-31 17:24:51 +00:00
Anton Korobeynikov
55998b56a0 Make StringMap to be more STL'ish. Patch by Mikhail Glushenkov!
llvm-svn: 46612
2008-01-31 12:30:40 +00:00
Anton Korobeynikov
45e5afd820 Whitespace cleanup
llvm-svn: 46611
2008-01-31 12:10:41 +00:00
Evan Cheng
705212577d Add an extra operand to LABEL nodes which distinguishes between debug, EH, or misc labels. This fixes the EH breakage. However I am not convinced this is *the* solution.
llvm-svn: 46609
2008-01-31 09:59:15 +00:00
Evan Cheng
a63f6736f3 MRegisterInfo::getLocation() is a really bad idea. Its function is to calculate the offset from frame pointer to a stack slot and then storing the delta in a MachineLocation object. The name is bad (it implies a getter), and MRegisterInfo doesn't need to know about MachineLocation.
Replace getLocation() with getFrameIndexOffset() which returns the delta from frame pointer to stack slot. Dwarf writer can then use the information for whatever it wants.

llvm-svn: 46597
2008-01-31 03:37:28 +00:00
Dan Gohman
562ed58408 Avoid unnecessarily casting away const, fixing a FIXME.
llvm-svn: 46591
2008-01-31 01:05:10 +00:00
Dan Gohman
703a070e12 Fix a typo in a comment.
llvm-svn: 46588
2008-01-31 00:44:33 +00:00
Dan Gohman
3993809a0c Rename ISD::FLT_ROUNDS to ISD::FLT_ROUNDS_ to avoid conflicting
with the real FLT_ROUNDS (defined in <float.h>).

llvm-svn: 46587
2008-01-31 00:41:03 +00:00
Dan Gohman
4326d513ab Create a new class, MemOperand, for describing memory references
in the backend. Introduce a new SDNode type, MemOperandSDNode, for
holding a MemOperand in the SelectionDAG IR, and add a MemOperand
list to MachineInstr, and code to manage them. Remove the offset
field from SrcValueSDNode; uses of SrcValueSDNode that were using
it are all all using MemOperandSDNode now.

Also, begin updating some getLoad and getStore calls to use the
PseudoSourceValue objects.

Most of this was written by Florian Brander, some
reorganization and updating to TOT by me.

llvm-svn: 46585
2008-01-31 00:25:39 +00:00
Evan Cheng
07f3cd3242 A semi-gross fix for a debug info issue. When inserting the "function start" label (i.e. first label in the entry block) take care to insert it at the beginning of the block.
llvm-svn: 46568
2008-01-30 19:35:32 +00:00
Evan Cheng
918b9c9335 Even though InsertAtEndOfBasicBlock is an ugly hack it still deserves a proper name. Rename it to EmitInstrWithCustomInserter since it does not necessarily insert
instruction at the end.

llvm-svn: 46562
2008-01-30 18:18:23 +00:00
Dan Gohman
dd9be7d05e Add a new PseudoSourceValue class, which will be used to help track
memory reference information in the backend. Most of this was written by
Florian Brander, cleanup and updating to TOT by me.

llvm-svn: 46556
2008-01-30 16:35:31 +00:00
Dan Gohman
13d1327796 Factor the addressing mode and the load/store VT out of LoadSDNode
and StoreSDNode into their common base class LSBaseSDNode. Member
functions getLoadedVT and getStoredVT are replaced with the common
getMemoryVT to simplify code that will handle both loads and stores.

llvm-svn: 46538
2008-01-30 00:15:11 +00:00