893 Commits

Author SHA1 Message Date
Evan Cheng
1fb8aedd1e Fix some significant problems with constant pools that resulted in unnecessary paddings between constant pool entries, larger than necessary alignments (e.g. 8 byte alignment for .literal4 sections), and potentially other issues.
1. ConstantPoolSDNode alignment field is log2 value of the alignment requirement. This is not consistent with other SDNode variants.
2. MachineConstantPool alignment field is also a log2 value.
3. However, some places are creating ConstantPoolSDNode with alignment value rather than log2 values. This creates entries with artificially large alignments, e.g. 256 for SSE vector values.
4. Constant pool entry offsets are computed when they are created. However, asm printer group them by sections. That means the offsets are no longer valid. However, asm printer uses them to determine size of padding between entries.
5. Asm printer uses expensive data structure multimap to track constant pool entries by sections.
6. Asm printer iterate over SmallPtrSet when it's emitting constant pool entries. This is non-deterministic.


Solutions:
1. ConstantPoolSDNode alignment field is changed to keep non-log2 value.
2. MachineConstantPool alignment field is also changed to keep non-log2 value.
3. Functions that create ConstantPool nodes are passing in non-log2 alignments.
4. MachineConstantPoolEntry no longer keeps an offset field. It's replaced with an alignment field. Offsets are not computed when constant pool entries are created. They are computed on the fly in asm printer and JIT.
5. Asm printer uses cheaper data structure to group constant pool entries.
6. Asm printer compute entry offsets after grouping is done.
7. Change JIT code to compute entry offsets on the fly.

llvm-svn: 66875
2009-03-13 07:51:59 +00:00
Bill Wendling
fa54bc2052 Oops...I committed too much.
llvm-svn: 66867
2009-03-13 04:39:26 +00:00
Bill Wendling
b02eadf660 Temporarily XFAIL this test.
llvm-svn: 66866
2009-03-13 04:37:11 +00:00
Bob Wilson
5b15d01ff3 Fix BuildVectorSDNode::isConstantSplat to handle one-element vectors.
It is an error to call APInt::zext with a size that is equal to the value's
current size, so use zextOrTrunc instead.

llvm-svn: 66039
2009-03-04 17:47:01 +00:00
Bob Wilson
85cefe8567 Generalize BuildVectorSDNode::isConstantSplat to use APInts and handle
arbitrary vector sizes.  Add an optional MinSplatBits parameter to specify
a minimum for the splat element size.  Update the PPC target to use the
revised interface.

llvm-svn: 65899
2009-03-02 23:24:16 +00:00
Bob Wilson
d8ea0e144e Combine PPC's GetConstantBuildVectorBits and isConstantSplat functions to a new
method in a BuildVectorSDNode "pseudo-class".

llvm-svn: 65747
2009-03-01 01:13:55 +00:00
Evan Cheng
a49de9de2e Revert BuildVectorSDNode related patches: 65426, 65427, and 65296.
llvm-svn: 65482
2009-02-25 22:49:59 +00:00
Scott Michel
e2fdc31759 Expand tabs to spaces (overlooked in previous commit)
llvm-svn: 65427
2009-02-25 03:57:49 +00:00
Scott Michel
bb878288cb Remove all "cached" data from BuildVectorSDNode, preferring to retrieve
results via reference parameters.

This patch also appears to fix Evan's reported problem supplied as a
reduced bugpoint test case.

llvm-svn: 65426
2009-02-25 03:12:50 +00:00
Dan Gohman
4f356bb9b0 Fix a ValueTracking rule: RHS means operand 1, not 0. Add a simple
ashr instcombine to help expose this code. And apply the fix to
SelectionDAG's copy of this code too.

llvm-svn: 65364
2009-02-24 02:00:40 +00:00
Scott Michel
9d31aca679 Introduce the BuildVectorSDNode class that encapsulates the ISD::BUILD_VECTOR
instruction. The class also consolidates the code for detecting constant
splats that's shared across PowerPC and the CellSPU backends (and might be
useful for other backends.) Also introduces SelectionDAG::getBUID_VECTOR() for
generating new BUILD_VECTOR nodes.

llvm-svn: 65296
2009-02-22 23:36:09 +00:00
Scott Michel
cf0da6c597 Remove trailing whitespace to reduce later commit patch noise.
(Note: Eventually, commits like this will be handled via a pre-commit hook that
 does this automagically, as well as expand tabs to spaces and look for 80-col
 violations.)

llvm-svn: 64827
2009-02-17 22:15:04 +00:00
Dale Johannesen
9c310711bb Use getDebugLoc forwarder instead of getNode()->getDebugLoc.
No functional change.

llvm-svn: 64026
2009-02-07 19:59:05 +00:00
Dale Johannesen
8ba7132128 Make SDNode constructors take a DebugLoc always.
Adjust derived classes to pass UnknownLoc where
a DebugLoc does not make sense.  Pick one of
DebugLoc and non-DebugLoc variants to survive
for all such classes.

llvm-svn: 64000
2009-02-07 02:15:05 +00:00
Dale Johannesen
a72d41a67b Remove now-unused constructors.
llvm-svn: 63995
2009-02-07 01:27:09 +00:00
Dale Johannesen
62fd95d6ec Get rid of the last non-DebugLoc versions of getNode!
Many targets build placeholder nodes for special operands, e.g.
GlobalBaseReg on X86 and PPC for the PIC base.  There's no
sensible way to associate debug info with these.  I've left
them built with getNode calls with explicit DebugLoc::getUnknownLoc operands. 
I'm not too happy about this but don't see a good improvement;
I considered adding a getPseudoOperand or something, but it
seems to me that'll just make it harder to read.

llvm-svn: 63992
2009-02-07 00:55:49 +00:00
Dale Johannesen
84935759d5 Remove more non-DebugLoc getNode variants. Use
getCALLSEQ_{END,START} to permit passing no DebugLoc
there.  UNDEF doesn't logically have DebugLoc; add
getUNDEF to encapsulate this.

llvm-svn: 63978
2009-02-06 23:05:02 +00:00
Dale Johannesen
400dc2e2e4 Remove more non-DebugLoc versions of getNode.
llvm-svn: 63969
2009-02-06 21:50:26 +00:00
Dale Johannesen
ab8e4425a3 Eliminate remaining non-DebugLoc version of getTargetNode.
llvm-svn: 63951
2009-02-06 19:16:40 +00:00
Dale Johannesen
2c4cf2752d get rid of some non-DebugLoc getTargetNode variants.
llvm-svn: 63909
2009-02-06 02:08:06 +00:00
Dale Johannesen
9f3f72f144 Get rid of one more non-DebugLoc getNode and
its corresponding getTargetNode.  Lots of
caller changes.

llvm-svn: 63904
2009-02-06 01:31:28 +00:00
Dale Johannesen
f80493bbfd Remove a non-DebugLoc version of getNode.
llvm-svn: 63889
2009-02-05 22:07:54 +00:00
Dale Johannesen
3eb373f5ce Remove 3 non-DebugLoc variants of getNode.
llvm-svn: 63886
2009-02-05 21:20:44 +00:00
Mon P Wang
3f0e0a6dea Fix a bug where we were not emitting a cvt rnd sat node for converting
between a unsigned integer and signed integer.

llvm-svn: 63831
2009-02-05 04:47:42 +00:00
Dale Johannesen
12c572b6fa Get rid of 3 non-DebugLoc getNode variants.
llvm-svn: 63808
2009-02-05 01:01:16 +00:00
Dale Johannesen
7ae8c8b108 Remove non-DebugLoc versions of getMergeValues, ZeroExtendInReg.
llvm-svn: 63800
2009-02-05 00:20:09 +00:00
Dale Johannesen
f08a47bb70 Remove non-DebugLoc forms of CopyToReg and CopyFromReg.
Adjust callers.

llvm-svn: 63789
2009-02-04 23:02:30 +00:00
Stuart Hastings
556bd92698 80 column rule.
llvm-svn: 63768
2009-02-04 20:30:10 +00:00
Dale Johannesen
021052a705 Remove non-DebugLoc versions of getLoad and getStore.
Adjust the many callers of those versions.

llvm-svn: 63767
2009-02-04 20:06:27 +00:00
Stuart Hastings
ffee3d831a Since I'm obliged to work with a development OS that currently doesn't
support GraphViz, I've been using the foo->dump() facility.  This
patch is a minor rewrite to the SelectionDAG dump() stuff to make it a
little more helpful.  The existing foo->dump() functionality does not
change; this patch adds foo->dumpr().  All of this is only useful when
running LLVM under a debugger.

llvm-svn: 63736
2009-02-04 16:46:19 +00:00
Dale Johannesen
679073b420 Remove non-DebugLoc forms of the exotic forms
of Lod and Sto; patch uses.

llvm-svn: 63716
2009-02-04 02:34:38 +00:00
Dale Johannesen
f2bb6f09a3 Remove some more non-DebugLoc versions of construction
functions, with callers adjusted to fit.

llvm-svn: 63705
2009-02-04 01:48:28 +00:00
Dale Johannesen
85263882aa Remove a few non-DebugLoc versions of node creation
functions.

llvm-svn: 63703
2009-02-04 01:17:06 +00:00
Dale Johannesen
3a09f5589d DebugLoc propagation; adjustment to things omitted
from SelectionDagBuild.

llvm-svn: 63680
2009-02-03 23:04:43 +00:00
Dale Johannesen
abf66b8343 Add some DL propagation to places that didn't
have it yet.  More coming.

llvm-svn: 63673
2009-02-03 22:26:09 +00:00
Devang Patel
70da8e8425 First initialize DAG otherwise dwarf writer is used uninitialized.
Duncan spotted this. Thanks!

llvm-svn: 63641
2009-02-03 18:46:32 +00:00
Dale Johannesen
db39362c90 Fill in some missing DL propagation in getNode()s.
llvm-svn: 63595
2009-02-03 01:55:44 +00:00
Dale Johannesen
f1163e9a4d Propagation in TargetLowering. Includes passing a DL
into SimplifySetCC which gets called elsewhere.

llvm-svn: 63583
2009-02-03 00:47:48 +00:00
Dan Gohman
76a07f59d4 Use the SubclassData field to hold ExtType, isTrunc, and MemIndexedMode
information. This eliminates the need for the Flags field in MemSDNode,
so this makes LoadSDNode and StoreSDNode smaller. Also, it makes
FoldingSetNodeIDs for loads and stores two AddIntegers smaller.

llvm-svn: 63577
2009-02-03 00:08:45 +00:00
Dale Johannesen
ae7992a333 Commit missing files.
llvm-svn: 63545
2009-02-02 20:47:48 +00:00
Duncan Sands
3ed768868d Fix PR3453 and probably a bunch of other potential
crashes or wrong code with codegen of large integers:
eliminate the legacy getIntegerVTBitMask and
getIntegerVTSignBit methods, which returned their
value as a uint64_t, so couldn't handle huge types.

llvm-svn: 63494
2009-02-01 18:06:53 +00:00
Duncan Sands
41826036b1 Fix PR3401: when using large integers, the type
returned by getShiftAmountTy may be too small
to hold shift values (it is an i8 on x86-32).
Before and during type legalization, use a large
but legal type for shift amounts: getPointerTy;
afterwards use getShiftAmountTy, fixing up any
shift amounts with a big type during operation
legalization.  Thanks to Dan for writing the
original patch (which I shamelessly pillaged).

llvm-svn: 63482
2009-01-31 15:50:11 +00:00
Mon P Wang
cf9ba82324 If unsafe FP optimization is not set, don't allow -(A-B) => B-A because
when A==B, -0.0 != +0.0.

llvm-svn: 63474
2009-01-31 06:07:45 +00:00
Dale Johannesen
555a375bb6 Make LowerCallTo and LowerArguments take a DebugLoc
argument.  Adjust all callers and overloaded versions.

llvm-svn: 63444
2009-01-30 23:10:59 +00:00
Bill Wendling
8fb81f1b3d Get rid of the non-DebugLoc-ified getNOT() method.
llvm-svn: 63442
2009-01-30 23:03:19 +00:00
Bill Wendling
c409318562 Propagate debug loc info for some of the *_EXTEND functions.
llvm-svn: 63434
2009-01-30 22:23:15 +00:00
Bill Wendling
cab9a2eef5 DebugLoc form of getNOT().
llvm-svn: 63433
2009-01-30 22:11:22 +00:00
Dan Gohman
ef04ed5477 In the case of an extractelement on an insertelement value,
the element indices may be equal if either one is not a
constant.

llvm-svn: 63311
2009-01-29 16:10:46 +00:00
Bill Wendling
a434d930ff Revert r63273. This was already implemented by Dale. There's no need for my
change.

llvm-svn: 63301
2009-01-29 09:01:55 +00:00
Bill Wendling
50338007b9 - Add DebugLoc to getTargetNode().
- Modify TableGen to add the DebugLoc when calling getTargetNode.

(The light-weight wrappers are only temporary. The non-DebugLoc version will be
removed once the whole debug info stuff is finished with.)

llvm-svn: 63273
2009-01-29 05:27:31 +00:00