96c1457bff
Teach if-conversion about instructions that were already predicated, e.g. conditional move.
...
llvm-svn: 37964
2007-07-06 23:24:39 +00:00
9d97420473
The exception handling intrinsics return values,
...
so must be lowered to a value, not nothing at all.
Subtle point: I made eh_selector return 0 and
eh_typeid_for return 1. This means that only
cleanups (destructors) will be run as the exception
unwinds [if eh_typeid_for returned 0 then it would
be as if the first catch always matched, and the
corresponding handler would be run], which is
probably want you want in the CBE.
llvm-svn: 37947
2007-07-06 14:46:23 +00:00
c5b1984f57
Indexes into the list of filter ids cannot be output
...
directly: they need to be turned into byte offsets
(often the same, but may not be if there are many
type infos).
llvm-svn: 37942
2007-07-06 12:46:24 +00:00
b567e3ffb0
Add the byval attribute
...
llvm-svn: 37940
2007-07-06 10:57:03 +00:00
003c0b1f90
Remove propagateEHRegister in favour of a more limited
...
fix, that is adequate while PR1508 remains unresolved.
llvm-svn: 37938
2007-07-06 09:18:59 +00:00
81df18a50a
Remove ExtractGlobalVariable - use StripPointerCasts
...
instead.
llvm-svn: 37937
2007-07-06 09:10:03 +00:00
fc7010d962
Workaround of getCopyToRegs and getCopyFromRegs bugs for big-endian machines.
...
llvm-svn: 37935
2007-07-06 01:47:35 +00:00
642be16bbf
Change CalculateHeights and CalculateDepths to be non-recursive.
...
llvm-svn: 37934
2007-07-06 01:37:28 +00:00
a282694acf
Make the debug string for ISD::MERGE_VALUES consistent with the others.
...
llvm-svn: 37922
2007-07-05 20:15:43 +00:00
d258e80583
Add a parameter to getCopyToParts and getCopyFromParts to specify whether
...
endian swapping should be done, and update the code to use it. This fixes
some register ordering issues on big-endian systems, such as PowerPC,
introduced by the recent illegal by-val arguments changes.
llvm-svn: 37921
2007-07-05 20:12:34 +00:00
e16561cd5d
Here is the bulk of the sanitizing.
...
Almost all occurrences of "bytecode" in the sources have been eliminated.
llvm-svn: 37913
2007-07-05 17:07:56 +00:00
4836e3a6f8
Make sure only one copy of a filter is placed in the
...
exception handling table if we encounter it multiple
times. Filters could be folded harder than this, but
that would mean a lot more work for not much gain.
llvm-svn: 37908
2007-07-05 15:15:01 +00:00
bb6ecf0067
Better assertion messages.
...
llvm-svn: 37890
2007-07-05 07:05:38 +00:00
fe80638417
Extend eh.selector to support both catches and filters.
...
Drop the eh.filter intrinsic.
llvm-svn: 37875
2007-07-04 20:52:51 +00:00
06563a8702
Fix several over-aggressive folds for undef nodes in dagcombine, to
...
follow the rules for undef used in instcombine.
llvm-svn: 37851
2007-07-03 14:03:57 +00:00
a2b3c175db
Fix for PR 1505 (and 1489). Rewrite X87 register
...
model to include f32 variants. Some factoring
improvments forthcoming.
llvm-svn: 37847
2007-07-03 00:53:03 +00:00
533dd16a7f
Replace ExpandScalarFormalArgs and ExpandScalarCallArgs with the newly
...
refactored getCopyFromParts and getCopyToParts, which are more general.
This effectively adds support for lowering illegal by-val vector call
arguments.
llvm-svn: 37843
2007-07-02 16:18:06 +00:00
9a70823375
Teach GetNegatedExpression to negate 0-B to B in UnsafeFPMath mode, and
...
visitFSUB to fold 0-B to -B in UnsafeFPMath mode. Also change visitFNEG
to use isNegatibleForFree/GetNegatedExpression instead of doing a subset
of the same thing manually.
This fixes test/CodeGen/X86/negative-sin.ll.
llvm-svn: 37842
2007-07-02 15:48:56 +00:00
fa68d069ad
Only do FNEG xform when the vector type is a floating point type.
...
llvm-svn: 37818
2007-06-29 21:44:35 +00:00
cf2a51e8db
Remove unused variables.
...
llvm-svn: 37816
2007-06-29 21:42:03 +00:00
9458e6a551
Fix a vector FP constant CSE bug.
...
llvm-svn: 37814
2007-06-29 21:36:04 +00:00
4c1e6f3804
Remove unnecessary attributions in comments.
...
llvm-svn: 37799
2007-06-29 03:42:23 +00:00
9468bfd932
Fix reference to cached end iterator invalidated by an erase operation.
...
Uncovered by _GLIBCXX_DEBUG.
llvm-svn: 37795
2007-06-29 02:49:11 +00:00
5b6f755575
Remove the "special tie breaker" because it resulted in inconsistent
...
ordering and thus violated the strict weak ordering requirement of
priority_queue. Uncovered by _GLIBCXX_DEBUG.
llvm-svn: 37794
2007-06-29 02:48:09 +00:00
451d1a6ecd
Fix misue of iterator pointing to erased object. Uncovered by
...
_GLIBCXX_DEBUG.
llvm-svn: 37793
2007-06-29 02:45:24 +00:00
0de7694de6
Fix an assertion failure in legalizing bitcast operators on targets where
...
vectors are split down to single elements as part of legalization.
llvm-svn: 37785
2007-06-29 00:09:08 +00:00
7867793aff
Add new TargetLowering code to provide the final register type that an
...
illegal value type will be transformed to, for code that needs the
register type after all transformations instead of just after the first
transformation.
Factor out the code that uses this information to do copy-from-regs and
copy-to-regs for various purposes into separate functions so that they
are done consistently.
llvm-svn: 37781
2007-06-28 23:29:44 +00:00
df0c705d7d
If a livein is not used in the block. It's live through.
...
llvm-svn: 37764
2007-06-27 18:47:28 +00:00
77f541ddfd
Partial fix for PR1502: If a EH register is needed in a successor of landing pad, add it as livein to all the blocks in the paths between the landing pad and the specified block.
...
llvm-svn: 37763
2007-06-27 18:45:32 +00:00
3b62d7265d
Rename ("shrinkify") MVT::isExtendedValueType to MVT::isExtendedVT.
...
llvm-svn: 37758
2007-06-27 16:08:04 +00:00
7139a48057
Use getVectorTypeBreakdown in FunctionLoweringInfo::CreateRegForValue
...
to compute the number and type of registers needed for vector values
instead of computing it manually. This fixes PR1529.
llvm-svn: 37755
2007-06-27 14:34:07 +00:00
f4e86da3a6
Make the comment for ScalarizeVectorOp mention that it is only for use
...
with single-element vectors.
llvm-svn: 37752
2007-06-27 14:06:22 +00:00
95759c0e78
ok, this is something of a dirty hack, but it seems to work. (fixes e.g.
...
the SPASS miscompilation)
llvm-svn: 37750
2007-06-27 09:01:14 +00:00
89183dcc14
ok, this much doesn't seem to bork anything
...
llvm-svn: 37749
2007-06-27 08:31:07 +00:00
a415dcaa2a
revert evan's fixes (and my doofusness) since they had a huge code
...
quality hit. will look at this soon.
llvm-svn: 37748
2007-06-27 08:11:59 +00:00
540d329542
pull evan's fixes - should help the nightly tester (but there are still
...
some issues)
llvm-svn: 37747
2007-06-27 07:07:13 +00:00
e66f822ecc
Replace std::set with SmallPtrSet.
...
llvm-svn: 37746
2007-06-27 05:23:00 +00:00
6cf1371456
Fix an obvious bug. Old code only worked for the entry block.
...
llvm-svn: 37743
2007-06-27 01:16:36 +00:00
4bf87f1f44
Correctly handle implcit def / use operands.
...
llvm-svn: 37740
2007-06-26 21:05:13 +00:00
d8417d9199
Properly handle kills of a physical register which has sub-registers that are read by later instructions.
...
llvm-svn: 37739
2007-06-26 21:03:35 +00:00
f4dc5b26e2
tidy this file up a bit
...
llvm-svn: 37725
2007-06-26 00:21:58 +00:00
77416383e8
A bunch of fixes to the BigBlock allocator improve compile-time by ~20%
...
and code quality by ~2% on my tests.
A big thank you to Roman Levenstein for this patch! See
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070618/050717.html
for more details.
llvm-svn: 37724
2007-06-25 23:46:54 +00:00
a866514528
Generalize MVT::ValueType and associated functions to be able to represent
...
extended vector types. Remove the special SDNode opcodes used for pre-legalize
vector operations, and the special MVT::Vector type used with them. Adjust
lowering and legalize to work with the normal SDNode kinds instead, and to
use the normal MVT functions to work with vector types instead of using the
two special operands that the pre-legalize nodes held.
This allows pre-legalize and post-legalize DAGs, and the code that operates
on them, to be more consistent. Pre-legalize vector operators can be handled
more consistently with scalar operators. And, -view-dag-combine1-dags and
-view-legalize-dags now look prettier for vector code.
llvm-svn: 37719
2007-06-25 16:23:39 +00:00
309d3d51b3
Move ComputeMaskedBits, MaskedValueIsZero, and ComputeNumSignBits from
...
TargetLowering to SelectionDAG so that they have more convenient
access to the current DAG, in preparation for the ValueType routines
being changed from standalone functions to members of SelectionDAG for
the pre-legalize vector type changes.
llvm-svn: 37704
2007-06-22 14:59:07 +00:00
81a752aa95
check in the BigBlock local register allocator
...
llvm-svn: 37703
2007-06-22 08:27:12 +00:00
e3c4419953
std::set is really really terrible. Switch to SmallPtrSet to reduce compile time. For Duraid's example. The overall isel time is reduced from 0.6255 sec to 0.1876 sec.
...
llvm-svn: 37701
2007-06-22 01:35:51 +00:00
8e8d34b220
Tidy up ValueType names in comments.
...
llvm-svn: 37688
2007-06-21 14:48:26 +00:00
04deef3a49
Rename TargetLowering::getNumElements and friends to
...
TargetLowering::getNumRegisters and similar, to avoid confusion with
the actual number of elements for vector types.
llvm-svn: 37687
2007-06-21 14:42:22 +00:00
aa5f5d960d
Xforms:
...
(add (select cc, 0, c), x) -> (select cc, x, (add, x, c))
(sub x, (select cc, 0, c)) -> (select cc, x, (sub, x, c))
llvm-svn: 37685
2007-06-21 07:39:16 +00:00
a955c02c9b
Avoid if-converting simple block that ends with unconditional branch or fallthrough unless it branches / falls to the 'false' block. Not profitable, may end up increasing code size.
...
llvm-svn: 37660
2007-06-19 21:45:13 +00:00
a7644dd9b9
Pass a SelectionDAG into SDNode::dump everywhere it's used, in prepration
...
for needing the DAG node to print pre-legalize extended value types, and
to get better debug messages with target-specific nodes.
llvm-svn: 37656
2007-06-19 14:13:56 +00:00
37228f6ce5
document and hide two options.
...
llvm-svn: 37651
2007-06-19 05:47:49 +00:00
9e82064924
Replace M_REMATERIALIZIBLE and the newly-added isOtherReMaterializableLoad
...
with a general target hook to identify rematerializable instructions. Some
instructions are only rematerializable with specific operands, such as loads
from constant pools, while others are always rematerializable. This hook
allows both to be identified as being rematerializable with the same
mechanism.
llvm-svn: 37644
2007-06-19 01:48:05 +00:00
a7ca624028
Replace TargetInstrInfo::CanBeDuplicated() with a M_NOT_DUPLICABLE bit.
...
llvm-svn: 37643
2007-06-19 01:26:51 +00:00
26be02febf
add isVarArg to CCState
...
llvm-svn: 37640
2007-06-19 00:11:09 +00:00
0598b2d5eb
Fix some fragile code wrt CFG edge updating.
...
llvm-svn: 37634
2007-06-18 22:44:57 +00:00
2afd702c20
Move CorrectExtraCFGEdges() from BranchFolding.cpp to a MachineBasicBlock method.
...
llvm-svn: 37633
2007-06-18 22:43:58 +00:00
e31adc8ab9
make ComputeTopDownOrdering significantly faster and use less stack space
...
by making it non-recursive
llvm-svn: 37629
2007-06-18 21:28:10 +00:00
8c7333266c
Make chain dependencies blue, in addition to being dashed.
...
llvm-svn: 37626
2007-06-18 15:30:16 +00:00
51eb2c3bbe
Properly remove duplicate instructions as result of diamond if-conversion. Other bug fixes.
...
llvm-svn: 37623
2007-06-18 08:37:25 +00:00
3a51c854d0
Really turn if-converter loose:
...
1. Consider all possible ifcvt cases at once. No longer restricted to bottom
up iterative approach.
2. Sort all possible cases based on a cost function. Perform the most profitable
ones first invalidate others that target the same blocks.
3. Fixed a number of bugs related to block duplication.
llvm-svn: 37613
2007-06-16 09:34:52 +00:00
e199f97fa8
Codegen support (stripped out) for the annotate attribute.
...
llvm-svn: 37608
2007-06-15 22:26:58 +00:00
23402fcaef
Not every predicable block can be safely duplicated.
...
llvm-svn: 37607
2007-06-15 21:18:05 +00:00
f852e339b6
Fix CodeGen/X86/inline-asm-x-scalar.ll:test4, by retaining regclass info
...
for tied register constraints.
llvm-svn: 37601
2007-06-15 19:11:01 +00:00
234a90e83e
MachineInstr::isPredicable() is no longer needed.
...
llvm-svn: 37599
2007-06-15 19:06:07 +00:00
92bf2c628c
Workaround for PR1508.
...
llvm-svn: 37597
2007-06-15 19:04:19 +00:00
cd5f5e3ecd
Extra edges are deleted later if needed.
...
llvm-svn: 37593
2007-06-15 17:34:48 +00:00
92fb5453c3
Allow small blocks to be duplicated to enable if-conversion.
...
llvm-svn: 37590
2007-06-15 07:36:12 +00:00
add977670f
No really, clear predcessors states.
...
llvm-svn: 37581
2007-06-14 23:34:09 +00:00
9fc56c079d
If BB is predicated, invalidate its predecessor(s) which would if-convert it. It needs to be re-analyzed.
...
llvm-svn: 37580
2007-06-14 23:13:19 +00:00
5c4413120f
Rename MVT::getVectorBaseType to MVT::getVectorElementType.
...
llvm-svn: 37579
2007-06-14 22:58:02 +00:00
e08f0eefa6
Fix typo.
...
llvm-svn: 37577
2007-06-14 21:26:08 +00:00
4a4a8eb00e
Add a target hook to allow loads from constant pools to be rematerialized, and an
...
implementation for x86.
llvm-svn: 37576
2007-06-14 20:50:44 +00:00
1e6f08b2a2
Fix some stupid bugs that have effectively disabled if-conversion.
...
llvm-svn: 37575
2007-06-14 20:28:52 +00:00
7413736a7e
Only correctly lower exception handing intrinsics if exception handling is
...
turned on. Likewise for scanning of invokes to mark landing pads.
llvm-svn: 37570
2007-06-13 16:53:21 +00:00
26455c4ae0
Introduce new SelectionDAG node opcodes VEXTRACT_SUBVECTOR and
...
VCONCAT_VECTORS. Use these for CopyToReg and CopyFromReg legalizing in
the case that the full register is to be split into subvectors instead
of scalars. This replaces uses of VBIT_CONVERT to present values as
vector-of-vector types in order to make whole subvectors accessible via
BUILD_VECTOR and EXTRACT_VECTOR_ELT.
This is in preparation for adding extended ValueType values, where
having vector-of-vector types is undesirable.
llvm-svn: 37569
2007-06-13 15:12:02 +00:00
cbd51c8b60
When creating CopyFromReg nodes, always use legal types. And use the
...
correct types for the result vector, even though it is currently bitcasted
to a different type immediately.
llvm-svn: 37568
2007-06-13 14:55:16 +00:00
97f7236e70
The fix that was applied for PR1224 stops the compiler
...
crashing but breaks exception handling. The problem
described in PR1224 is that invoke is a terminator that
can produce a value. The value may be needed in other
blocks. The code that writes to registers values needed
in other blocks runs before terminators are lowered (in
this case invoke) so asserted because the value was not
yet available. The fix that was applied was to do invoke
lowering earlier, before writing values to registers.
The problem this causes is that the code to copy values
to registers can be output after the invoke call. If
an exception is raised and control is passed to the
landing pad then this copy-code will never execute. If
the value is needed in some code path reached via the
landing pad then that code will get something bogus.
So revert the original fix and simply skip invoke values
in the general copying to registers code. Instead copy
the invoke value to a register in the invoke lowering code.
llvm-svn: 37567
2007-06-13 05:51:31 +00:00
cbaaff58bf
Typo
...
llvm-svn: 37566
2007-06-13 00:04:00 +00:00
9acfa7b063
Now if-converting all 4 variants of triangles.
...
llvm-svn: 37565
2007-06-12 23:54:05 +00:00
4c2f003971
Fix a typo in bswap lowering.
...
llvm-svn: 37544
2007-06-11 23:16:16 +00:00
4dd31a7b1b
Restructure code to reduce ifcvt compile time cost.
...
llvm-svn: 37543
2007-06-11 22:26:22 +00:00
14b62a553e
Fix the build.
...
llvm-svn: 37537
2007-06-10 00:19:17 +00:00
2117d1f20e
Don't change CFG during analysis stage. Do so during ifcvt and invalidate predecessors accordingly.
...
llvm-svn: 37531
2007-06-09 01:03:43 +00:00
288f1545a6
Carefully remove extraneous CFG edges after each ifcvt.
...
llvm-svn: 37529
2007-06-08 22:01:07 +00:00
df1a42935b
Correct transfer predicate information.
...
llvm-svn: 37524
2007-06-08 19:17:12 +00:00
e93ccc013f
Hidden options to help debugging ifcvt issues.
...
llvm-svn: 37523
2007-06-08 19:10:51 +00:00
02f6e9b621
Factor live variable analysis so it does not do register coalescing
...
simultaneously. Move that pass to SimpleRegisterCoalescing.
This makes it easier to implement alternative register allocation and
coalescing strategies while maintaining reuse of the existing live
interval analysis.
llvm-svn: 37520
2007-06-08 17:18:56 +00:00
7783f82e21
Allow more cmp / bcc to be predicated; clean up triangle ifcvt checking code.
...
llvm-svn: 37518
2007-06-08 09:36:04 +00:00
dc8fc1550e
Use more realistically sized vectors. Reserve capacity if we know in advance
...
how much will be used.
llvm-svn: 37515
2007-06-08 08:59:11 +00:00
86798e5e11
Make throttle a hidden parameter, per review.
...
llvm-svn: 37511
2007-06-08 01:08:52 +00:00
52fcf022f7
Throttle tail merging; handling blocks with large numbers of predecessors
...
is too slow.
llvm-svn: 37509
2007-06-08 00:34:27 +00:00
1236ef7bcc
Only remove the edge from entry to false if false block is merged.
...
llvm-svn: 37503
2007-06-07 22:31:28 +00:00
9a4d987a5f
Do not change the size of function arguments. PR 1489.
...
llvm-svn: 37496
2007-06-07 21:07:15 +00:00
d3f3f0adad
ifcvt a triangle: don't merge ifcvt block with rejoin block if it can fall through to it. If merged, the resulting block is not a candidate for iterative ifcvting since it contains both predicated and non-predicated code.
...
llvm-svn: 37487
2007-06-07 08:13:00 +00:00
be9859eea2
Lots of bug fixes. Now finally in a reasonable state.
...
llvm-svn: 37485
2007-06-07 02:12:15 +00:00
cfb6f40424
Quick patch to fix the build, based on what it appears Evan meant to write.
...
Evan, please check that this is in fact correct.
llvm-svn: 37471
2007-06-06 16:22:00 +00:00
29d1dc6f1a
Fold the exception actions table harder: if two typeid lists start the
...
same, only output one copy of the common part.
llvm-svn: 37470
2007-06-06 15:37:31 +00:00
9030b98aca
Lots of bug fixes.
...
llvm-svn: 37467
2007-06-06 10:16:17 +00:00
61166501a1
Additional fix for PR1422: make sure the landing pad label is placed in the
...
correct machine basic block - do not rely on the eh.exception intrinsic
being in the landing pad: the loop optimizers can move it out.
llvm-svn: 37463
2007-06-06 10:05:18 +00:00
e4ec918be0
If a unconditional branch is added to branch to the false path during ifcvt, the predicated block cannot be iteratively ifcvted.
...
llvm-svn: 37456
2007-06-06 02:08:52 +00:00
b30a89457c
Minor statistics counting bug.
...
llvm-svn: 37451
2007-06-06 01:12:44 +00:00
30565998bb
Fix a couple of typos and be smarter about order of blocks when ifcvt a diamond.
...
llvm-svn: 37449
2007-06-06 00:57:55 +00:00
7948422b78
Fix diamond shape ifcvt bugs.
...
llvm-svn: 37444
2007-06-05 23:46:14 +00:00
c1a0b8ce1d
ReplaceUsesOfBlockWith() can modify the predecessors list.
...
llvm-svn: 37441
2007-06-05 22:03:53 +00:00
3e5bf0827c
Do not ifcvt if either true / false path is a backedge. Not profitable in almost all cases.
...
llvm-svn: 37440
2007-06-05 20:38:42 +00:00
2c1acd6d9e
I had a senior moment.
...
llvm-svn: 37433
2007-06-05 07:05:25 +00:00
6e4babe8cc
If the predicated block requires an early exit, end the block there and add a unconditional branch to false block. AnalyzeBranch() does not understand early exits.
...
llvm-svn: 37430
2007-06-05 01:31:40 +00:00
17aad8164e
Fix some subtle bugs: bug during succeessor copying; incorrectly updating states of ifcvted blocks.
...
llvm-svn: 37429
2007-06-05 00:07:37 +00:00
0558dda319
Tail merging wasn't working for predecessors of landing pads. PR 1496.
...
llvm-svn: 37427
2007-06-04 23:52:54 +00:00
53ce7de03b
Global ctors / dtors alignment shouldn't be hard-coded at 4. e.g. It could be 8 for 64-bit targets.
...
llvm-svn: 37421
2007-06-04 20:39:18 +00:00
91233153bf
Forgot to check for if iterator reached the end.
...
llvm-svn: 37420
2007-06-04 20:33:36 +00:00
b4c2690446
Pass the DAG to SDNode::dump to let it do more detailed dumps in some cases.
...
llvm-svn: 37413
2007-06-04 16:17:33 +00:00
92a7f3a65e
Resolve implicit alignment before computing the FoldingSet information so
...
that the CSE map always contains explicit alignment information. This allows
more loads to be CSE'd when there is a mix of explicit-alignment loads and
implicit-alignment loads.
Also, in SelectionDAG::FindModifiedNodeSlot, add the operands to the
FoldingSetNodeID before the load/store information instead of after, so
that it matches what is done elsewhere.
llvm-svn: 37411
2007-06-04 15:49:41 +00:00
312b723af2
Let IfConverter loose. Allow more aggressive subsumptions; reorder basic blocks to expose more ifcvt opportunities; code clean up and fixes.
...
llvm-svn: 37409
2007-06-04 06:47:22 +00:00
df75785594
Move ReplaceUsesOfBlockWith() out of BranchFolding into a MachineBasicBlock general facility.
...
llvm-svn: 37408
2007-06-04 06:44:01 +00:00
f708f73a1b
The semantics of invoke require that we always jump to the unwind block
...
(landing pad) when an exception unwinds through the call. This doesn't
quite match the way the dwarf unwinder works: by default it only jumps to
the landing pad if the catch or filter specification matches, and otherwise
it keeps on unwinding. There are two ways of specifying to the unwinder
that it should "always" (more on why there are quotes here later) jump to
the landing pad: follow the specification by a 0 typeid, or follow it by
the typeid for the NULL typeinfo. GCC does the first, and this patch makes
LLVM do the same as gcc. However there is a problem: the unwinder performs
optimizations based on C++ semantics (it only expects destructors to be
run if the 0 typeid fires - known as "cleanups"), meaning it assumes that no
exceptions will be raised and that the raised exception will be reraised
at the end of the cleanup code. So if someone writes their own LLVM code
using the exception intrinsics they will get a nasty surprise if they don't
follow these rules. The other possibility of using the typeid corresponding
to NULL (catch-all) causes the unwinder to make no assumptions, so this is
probably what we should use in the long-run. However since we are still
having trouble getting exception handling working properly, for the moment
it seems best to closely imitate GCC.
llvm-svn: 37399
2007-06-02 17:16:06 +00:00
c063f5f362
Integrate exception filter support and exception catch support. This
...
simplifies the code in DwarfWriter, allows for multiple filters and
makes it trivial to specify filters accompanied by cleanups or catch-all
specifications (see next patch). What a deal! Patch blessed by Anton.
llvm-svn: 37398
2007-06-02 16:53:42 +00:00
0a0ae932ca
Make LowerCTPOP() support arbitrary bitwidth integer type.
...
llvm-svn: 37397
2007-06-02 04:10:33 +00:00
6e7cdce773
Fix CorrectExtraCFGEdges to allow for multiple LandingPad targets.
...
llvm-svn: 37394
2007-06-02 00:08:15 +00:00
9746e3a22b
Fancier algorithm in tail-merge comment implemented, so remove comment.
...
llvm-svn: 37393
2007-06-01 23:04:28 +00:00
3c0a13762d
Implement smarter algorithm for choosing which blocks to tail-merge.
...
See test/CodeGen/X86/test-pic-jtbl.ll for a case where it works well;
shaves another 10K off our favorite benchmark. I was hesitant about
this because of compile speed, but seems to do OK on a bootstrap.
llvm-svn: 37392
2007-06-01 23:02:45 +00:00
4dcf1e8582
Correctly mark early-exit on the false path.
...
llvm-svn: 37387
2007-06-01 20:29:21 +00:00
706421e712
Since TypeInfos are passed as i8 pointers, a NULL TypeInfo should be passed
...
as a null i8 pointer not as a 0 i32.
llvm-svn: 37383
2007-06-01 08:18:30 +00:00
6a2cf070cc
Ifcvt triangle: don't ifcvt 'true' BB if it has other predecessors; don't merge 'false' BB if it has other predecessors.
...
llvm-svn: 37382
2007-06-01 07:41:07 +00:00
95c7917d92
Remove a bogus check. Even terminators in a ifcvt need to be predicated. Unconditional branches can usually be converted to conditional ones.
...
llvm-svn: 37380
2007-06-01 00:55:26 +00:00
20e05997f5
Allow multiple ifcvt candidates to share children blocks; add some debugging code.
...
llvm-svn: 37379
2007-06-01 00:12:12 +00:00
1a401e68a8
Arrange for only 1 of multiple branches to landing pad to be kept.
...
Do not remove empty landing pads (EH table needs to be updated)
llvm-svn: 37375
2007-05-31 21:54:00 +00:00
e6ccb6c5ed
Fix a typo.
...
llvm-svn: 37374
2007-05-31 20:53:33 +00:00
3e3ff30aa2
Fix the asmprinter so that a globalvalue can specify an explicit alignment
...
smaller than the preferred alignment, but so that the target can actually
specify a minimum alignment if needed. This fixes some objc protocol
failures Devang tracked down.
llvm-svn: 37373
2007-05-31 18:57:45 +00:00
5b0757a401
Fix PR1424.
...
When a function has FP, the register scavenging spill slot offset already
was calculated.
llvm-svn: 37371
2007-05-31 18:27:58 +00:00
905a8f4940
Change traversal order to bottom up in preparation for more aggressive if-conversion.
...
llvm-svn: 37365
2007-05-30 19:49:19 +00:00
397c4d9ef6
Fix CodeGen/PowerPC/2007-05-30-dagcombine-miscomp.ll, and PR1473.
...
llvm-svn: 37362
2007-05-30 16:30:06 +00:00
d14ad078c6
Changed per review comment.
...
llvm-svn: 37355
2007-05-30 00:32:01 +00:00
a69ebdbebc
Make stable_sort in tail merging actually be stable (it never was, but didn't
...
matter until my last change). Reenable tail merging by default.
llvm-svn: 37354
2007-05-29 23:47:50 +00:00
20f7d30f92
Don't merge in tail block of a diamond if it has more than one predecessors after if-conversion.
...
llvm-svn: 37353
2007-05-29 23:37:20 +00:00
c2237ce217
If there is an empty block between a source and its successor block, it still requires a unconditional branch.
...
llvm-svn: 37344
2007-05-29 22:31:16 +00:00
5983bdbb2c
Add missing const qualifiers.
...
llvm-svn: 37341
2007-05-29 18:35:22 +00:00
d7dc1ed64c
Correct the logic in LowerPartSet which cleared the bits from 0 to low-1.
...
llvm-svn: 37331
2007-05-26 03:43:13 +00:00
4698083b96
tighten up recursion depth again
...
llvm-svn: 37330
2007-05-25 02:19:06 +00:00
13f5f7df95
Silly boog.
...
llvm-svn: 37328
2007-05-25 00:59:01 +00:00
1409b6a59b
Blocks that cond-br and uncond-br/fallthrough to same block should have
...
only one successor.
llvm-svn: 37324
2007-05-24 18:31:55 +00:00
1af8c870c7
Fix for PR1444: do not create two successors to the same block.
...
Temporarily, this breaks CodeGen/Generic/2006-02-12-InsertLibraryCall.ll
by exposing an unrelated latent problem; working on that.
llvm-svn: 37323
2007-05-24 17:39:32 +00:00
30978078bf
Minor comment cleanups.
...
llvm-svn: 37321
2007-05-24 14:36:04 +00:00
703e0f8608
Add explicit qualification for namespace MVT members.
...
llvm-svn: 37320
2007-05-24 14:33:05 +00:00
a4d187b8ce
Fix a typo that caused combiner to create mal-formed pre-indexed store where value store is the same as the base pointer.
...
llvm-svn: 37318
2007-05-24 02:35:39 +00:00
f4a77d2481
Two tail merging improvements:
...
When considering blocks with more than 2 predecessors, merge the block with
the largest number of matching insns, rather than the first block found.
Considering that 1 matching insn is enough to show a win for candidates that
already end with a branch.
llvm-svn: 37315
2007-05-23 21:07:20 +00:00
3b327826db
Mark all calls as "could throw", when exceptions are enabled. Emit necessary LP info too. This fixes PR1439
...
llvm-svn: 37311
2007-05-23 11:08:31 +00:00
6509c0673f
prevent exponential recursion in isNegatibleForFree
...
llvm-svn: 37310
2007-05-23 07:35:22 +00:00
d0e669199b
Preliminary iterative if-conversion support.
...
llvm-svn: 37309
2007-05-23 07:23:16 +00:00
f9cbdc676c
name change requested by review of previous patch
...
llvm-svn: 37289
2007-05-22 18:31:04 +00:00
0ae7eb5e7c
Silence a warning.
...
llvm-svn: 37288
2007-05-22 18:13:40 +00:00
82810c8a13
Make tail merging the default, except on powerPC. There was no prior art
...
for a target-dependent default with a command-line override; this way
should be generally usable.
llvm-svn: 37285
2007-05-22 17:14:46 +00:00
e26c0916a3
If-convert early exit blocks (returns, etc.); bug fixes, etc.
...
llvm-svn: 37270
2007-05-21 22:22:58 +00:00
34e82a4508
Only emit one entry in the exception action table for each action, even if
...
it occurs for multiple landing pads.
llvm-svn: 37267
2007-05-21 18:50:28 +00:00
1fa8276e70
same patch as the previous one, but the symmetric case
...
llvm-svn: 37249
2007-05-19 00:46:51 +00:00
b08cbbd737
Disable the (A == (B-A)) -> 2*A == B xform when the sub has multiple uses (in
...
this case, the xform introduces an extra operation). This compiles
PowerPC/compare-duplicate.ll into:
_test:
subf r2, r3, r4
cmplw cr0, r2, r3
bne cr0, LBB1_2 ;F
instead of:
_test:
slwi r2, r3, 1
subf r3, r3, r4
cmplw cr0, r4, r2
bne cr0, LBB1_2 ;F
This is target independent of course.
llvm-svn: 37246
2007-05-19 00:43:44 +00:00
018cffbca4
Clean up.
...
llvm-svn: 37237
2007-05-18 19:32:08 +00:00
faaf716540
Change to depth-first traversal.
...
llvm-svn: 37236
2007-05-18 19:26:33 +00:00
dafda82755
Document an inefficiency in tail merging.
...
llvm-svn: 37235
2007-05-18 18:46:40 +00:00
b539df3389
Qualify calls to getTypeForValueType with MVT:: too.
...
llvm-svn: 37233
2007-05-18 18:41:29 +00:00
2e82cefd24
Some restructuring in preparation for most aggressive if-conversion.
...
llvm-svn: 37231
2007-05-18 18:14:37 +00:00
1796f1f8e9
Qualify several calls to functions in the MVT namespace, for consistency.
...
llvm-svn: 37230
2007-05-18 17:52:13 +00:00
f25d3a5d73
Watch out for blocks that end with a return.
...
llvm-svn: 37227
2007-05-18 17:06:53 +00:00
478b805956
If true / false blocks fallthrough before ifcvt, add unconditional branches to ifcvt'd block.
...
llvm-svn: 37200
2007-05-18 01:55:58 +00:00
f8956178af
Remove some unneeded branches. (spotted by Evan, thanks)
...
llvm-svn: 37198
2007-05-18 01:28:58 +00:00
0f745da4fe
Make use of target specific block size limits; bug fixes.
...
llvm-svn: 37195
2007-05-18 00:20:58 +00:00
a92b2b38ff
Move isSuccessor() offline, change it to use std::find.
...
llvm-svn: 37190
2007-05-17 23:58:53 +00:00
0184f88deb
disable MaskedValueIsZero, ComputeMaskedBits, and SimplifyDemandedBits for
...
i128 integers. The 64-bit masks are not wide enough to represent the results.
These should be converted to APInt someday.
llvm-svn: 37169
2007-05-17 18:19:23 +00:00
2135bc08d6
add expand support for ADDC/SUBC/ADDE/SUBE so we can codegen 128-bit add/sub on 32-bit (or less) targets
...
llvm-svn: 37168
2007-05-17 18:15:41 +00:00
429178d727
Add target hook to specify block size limit for if-conversion.
...
llvm-svn: 37134
2007-05-16 23:45:53 +00:00
7a6c175e7a
Don't fold bitconvert(load) for preinc/postdec loads. Likewise stores.
...
llvm-svn: 37130
2007-05-16 22:45:30 +00:00
af71610429
isBlockPredicable() always ignore terminal instructions; add comments.
...
llvm-svn: 37126
2007-05-16 21:54:37 +00:00
5ea933a009
Rename M_PREDICATED to M_PREDICABLE; Moved isPredicable() to MachineInstr.
...
llvm-svn: 37121
2007-05-16 20:56:08 +00:00
59ae77486c
Output exception call-sites in address order, as required by the unwinding
...
runtime.
llvm-svn: 37104
2007-05-16 12:12:23 +00:00
48fb92f75d
Use a ptr set instead of a linear search to unique TokenFactor operands.
...
This fixes PR1423
llvm-svn: 37102
2007-05-16 06:37:59 +00:00
35fc2119b0
Devang points out that we need an assertion here.
...
llvm-svn: 37097
2007-05-16 05:11:10 +00:00
288f133c71
Bug fix: should check ABI alignment, not pref. alignment.
...
llvm-svn: 37094
2007-05-16 02:04:50 +00:00
f5e53a58db
Initial commit of (very basic) if converter.
...
llvm-svn: 37092
2007-05-16 02:00:57 +00:00
420a85d0cf
Remove extra CFG edges before doing these passes; it makes them happier.
...
llvm-svn: 37089
2007-05-15 21:19:17 +00:00
3f142cbca2
Fix an infinite recursion in GetNegatedExpression.
...
llvm-svn: 37086
2007-05-15 17:05:43 +00:00
750e8c5d4d
The index into the actions table is a ULEB128 not a SLEB128.
...
llvm-svn: 37084
2007-05-15 13:54:14 +00:00
daed139420
Un-brain-dead-ify the lowering of part set for the reverse case.
...
llvm-svn: 37071
2007-05-15 02:26:52 +00:00
c7596efdad
Fix some subtle issues handling immediate values. This fixes
...
test/CodeGen/ARM/2007-05-14-InlineAsmCstCrash.ll
llvm-svn: 37069
2007-05-15 01:33:58 +00:00
4d728b0419
Added getNumExplicitOperands and findFirstPredOperand.
...
llvm-svn: 37064
2007-05-15 01:26:09 +00:00
e49c974a7c
implement a simple fneg optimization/propagation thing. This compiles:
...
CodeGen/PowerPC/fneg.ll into:
_t4:
fmul f0, f3, f4
fmadd f1, f1, f2, f0
blr
instead of:
_t4:
fneg f0, f3
fmul f0, f0, f4
fmsub f1, f1, f2, f0
blr
llvm-svn: 37054
2007-05-14 22:04:50 +00:00
e1595b6859
Only worry about intervening kill if there are more than one live ranges in the interval.
...
llvm-svn: 37052
2007-05-14 21:23:51 +00:00
c690cba7d9
Fix for PR1406:
...
v1 =
r2 = move v1
= op r2<kill>
...
r2 = move v1
= op r2<kill>
Clear the first r2 kill if v1 and r2 are joined.
llvm-svn: 37050
2007-05-14 21:10:05 +00:00
fc2377d4ed
When marking a register as being implicitly defined, make sure to clear its partial use info as well.
...
llvm-svn: 37046
2007-05-14 20:39:18 +00:00
764ae2a21b
Give names to the final result values of the part_set computations. This
...
just aids in readability and debugability of the output. No functional change.
llvm-svn: 37037
2007-05-14 17:21:17 +00:00
1ee0c8d563
Emit function debug frames in one atom. This will prevent us from generating incorrect assembler in case of both
...
debug information & exception information presented.
llvm-svn: 37019
2007-05-13 17:30:11 +00:00
bbaf55448b
Emit multiple common EH frames for multiple (including blank) personality
...
functions. This partly fixes PR1414: now we're restricted only to one
personality function per eh frame, not per module. Further work on
"multiple personalities" topic needs representative example.
llvm-svn: 37018
2007-05-13 15:42:26 +00:00
13da17843c
More DWARF-related things cleanup:
...
1. Fix PR1380
2. Apply Duncan's patch from PR1410
3. Insert workaround for "one personality function per module" as noted in PR1414
4. Emit correct debug frames for x86/linux. This partly fixes DebugInfo/2006-11-06-StackTrace.cpp: stack trace is
shown correctly, but arguments for function on top of stack are displayed incorrectly.
llvm-svn: 37015
2007-05-12 22:36:25 +00:00
bf283708f7
Get the size of auto arrays right, regardless of its changing size.
...
llvm-svn: 37006
2007-05-12 11:07:40 +00:00
79a71ec3ad
Fix http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070507/049516.html
...
llvm-svn: 36998
2007-05-11 23:14:43 +00:00
67286b1266
Perform correct actions numbers/sizes computation
...
llvm-svn: 36988
2007-05-11 08:47:35 +00:00
00d02442b0
Fix action No calculation in multiple-invoke-one-LP mode
...
llvm-svn: 36987
2007-05-11 08:23:57 +00:00
cc8f571bc8
Do not generate branches to entry block. This fixes several test suite
...
failures on PPC (can happen only when prologue code is null)
llvm-svn: 36979
2007-05-10 23:59:23 +00:00
ee02c7d2fb
Ooops. Some debugging stuff :)
...
llvm-svn: 36978
2007-05-10 22:38:46 +00:00
96142de3f0
Allow multiple invokes per landing pad. This (probably) fixes PR1410.
...
llvm-svn: 36977
2007-05-10 22:34:59 +00:00
b43fe52136
Later computations assume we are aligned at this point.
...
llvm-svn: 36975
2007-05-10 18:40:24 +00:00
ed5dad4306
TypeIds are indexed by j, not i
...
llvm-svn: 36974
2007-05-10 15:10:34 +00:00
6e16d09252
Make tail merging handle many more cases (all it can, I think).
...
llvm-svn: 36966
2007-05-10 01:01:49 +00:00
f325c2a65e
Can't fold the bit_convert is the store is a truncating store.
...
llvm-svn: 36962
2007-05-09 21:49:47 +00:00
192d09c2d9
Do not assert, when case range split metric is zero and JTs are not allowed: just emit binary tree in this case. This
...
fixes PR1403.
llvm-svn: 36959
2007-05-09 20:07:08 +00:00
31fd60ba0c
Change names from RA to something unique to get rid of naming conflicts with
...
certain linkers...
llvm-svn: 36944
2007-05-08 19:02:46 +00:00
9e17872c1d
Eliminate MarkVirtRegAliveInBlock recursion.
...
llvm-svn: 36943
2007-05-08 19:00:00 +00:00
562e45692e
Forgot a check.
...
llvm-svn: 36910
2007-05-07 21:36:06 +00:00
a4cf58a103
Enable a couple of xforms:
...
- (store (bitconvert v)) -> (store v) if resultant store does not require
higher alignment
- (bitconvert (load v)) -> (load (bitconvert*)v) if resultant load does not
require higher alignment
llvm-svn: 36908
2007-05-07 21:27:48 +00:00
9a25b3afcd
Handle some non-exit blocks in tail merging.
...
llvm-svn: 36907
2007-05-07 20:57:21 +00:00
671e8c4444
Parameter attributes on invoke calls were being lost due to the wrong
...
attribute index being used. Fix proposed by Anton Korobeynikov, who
asked me to implement and commit it for him. This is PR1398.
llvm-svn: 36906
2007-05-07 20:49:28 +00:00
a8fd7fdc25
Detabify
...
llvm-svn: 36891
2007-05-06 20:14:21 +00:00
e7da2d6ac3
Fix typo in comment.
...
llvm-svn: 36873
2007-05-06 13:37:16 +00:00
00282a21a5
Use the personality function that was registered with MMI rather than
...
hardwiring in the C++ one.
llvm-svn: 36789
2007-05-05 20:27:00 +00:00
07e6f3257c
Propagate alignment/volatility in two places.
...
Implement support for expanding a bitcast from an illegal vector type to
a legal one (e.g. 4xi32 -> 4xf32 in SSE1). This fixes PR1371 and
CodeGen/X86/2007-05-05-VecCastExpand.ll
llvm-svn: 36787
2007-05-05 19:39:05 +00:00
02528f5425
Spelling fix.
...
llvm-svn: 36781
2007-05-05 16:32:57 +00:00
4db0090339
Emit sections/directives in the proper order. This fixes PR1376. Also,
...
some small cleanup was made.
llvm-svn: 36780
2007-05-05 09:04:50 +00:00
4cb9eb81ef
A bitcast of a global variable may have been constant folded to a GEP -
...
handle this case too.
llvm-svn: 36745
2007-05-04 17:12:26 +00:00
044a0a8cfb
Don't create indexed load / store with zero offset!
...
llvm-svn: 36716
2007-05-03 23:52:19 +00:00
44a2ed66b1
Allow i/s to match (gv+c). This fixes CodeGen/PowerPC/2007-05-03-InlineAsm-S-Constraint.ll
...
and PR1382
llvm-svn: 36672
2007-05-03 16:54:34 +00:00
8c78a0bff0
Drop 'const'
...
llvm-svn: 36662
2007-05-03 01:11:54 +00:00
11940fbba3
Properly set arguments bitwidth of EHSELECT node
...
llvm-svn: 36654
2007-05-02 22:15:48 +00:00
e95c6ad802
Use 'static const char' instead of 'static const int'.
...
Due to darwin gcc bug, one version of darwin linker coalesces
static const int, which defauts PassID based pass identification.
llvm-svn: 36652
2007-05-02 21:39:20 +00:00
41223586a2
Fix build error.
...
llvm-svn: 36648
2007-05-02 20:37:47 +00:00
b538f67b1a
Fix couple of bugs connected with eh info:
...
1. Correct output offsets on Linux
2. Fix "style" of personality function. It shouldn't be indirect.
llvm-svn: 36633
2007-05-01 22:23:12 +00:00
09f162ca6a
Do not use typeinfo to identify pass in pass manager.
...
llvm-svn: 36632
2007-05-01 21:15:47 +00:00
8cd28f0fb1
If call frame is not part of stack frame and no dynamic alloc, eliminateFrameIndex() must adjust SP offset with size of call frames.
...
llvm-svn: 36625
2007-05-01 09:01:42 +00:00
b68343cdd8
Forgot about chain result; also UNDEF cannot have multiple values.
...
llvm-svn: 36622
2007-05-01 08:53:39 +00:00
27a625a74b
llvm bug #1350 , parts 1, 2, and 3.
...
llvm-svn: 36618
2007-05-01 05:57:02 +00:00
77c545e6b8
Under normal circumstances, when a frame pointer is not required, we reserve
...
argument space for call sites in the function immediately on entry to the
current function. This eliminates the need for add/sub sp brackets around call
sites. However, this is not always a good idea. If the "call frame" is large and
the target load / store instructions have small immediate field to encode sp
offset, this can cause poor codegen. In the worst case, this can make it
impossible to scavenge a register if the reserved spill slot is pushed too far
apart from sp / fp.
llvm-svn: 36607
2007-05-01 00:52:08 +00:00
a684cd23a5
* Only turn a load to UNDEF if all of its outputs have no uses (indexed loads
...
produce two results.)
* Do not touch volatile loads.
llvm-svn: 36604
2007-05-01 00:38:21 +00:00
4dbbace4ff
Fix PR1228 and CodeGen/Generic/2007-04-30-LandingPadBranchFolding.ll
...
llvm-svn: 36602
2007-04-30 23:35:00 +00:00
f90c8fbd2b
print isLandingPad() for MBBs
...
llvm-svn: 36600
2007-04-30 23:12:53 +00:00
8cfd33b647
Continue refactoring inline asm code. If there is an earlyclobber output
...
register, preallocate all input registers and the early clobbered output.
This fixes PR1357 and CodeGen/PowerPC/2007-04-30-InlineAsmEarlyClobber.ll
llvm-svn: 36599
2007-04-30 21:11:17 +00:00
6b77c3ed52
Updates.
...
llvm-svn: 36594
2007-04-30 18:42:09 +00:00
4333f8b1cf
refactor GetRegistersForValue to take OpInfo as an argument instead of various
...
pieces of it. No functionality change.
llvm-svn: 36592
2007-04-30 17:29:31 +00:00
ef07332504
refactor some code, no functionality change
...
llvm-svn: 36590
2007-04-30 17:16:27 +00:00
b58f93f273
Clean up multi-line asam string printing. Instead of printing:
...
# InlineAsm Start
subfc r3,r5,r4
subfze r4,r3
# InlineAsm End
print:
# InlineAsm Start
subfc r3,r5,r4
subfze r4,r3
# InlineAsm End
llvm-svn: 36589
2007-04-30 17:00:18 +00:00
412d61af43
generalize aggregate handling
...
llvm-svn: 36568
2007-04-29 18:58:03 +00:00
546ea7ea88
Implement review feedback
...
llvm-svn: 36564
2007-04-29 18:02:48 +00:00
401d8db381
memory operands that have a direct operand should have their stores created
...
before the copies into physregs are done. This avoids having flag operands
skip the store, causing cycles in the dag at sched time. This fixes infinite
loops on these tests:
test/CodeGen/Generic/2007-04-08-MultipleFrameIndices.ll for PR1308
test/CodeGen/PowerPC/2007-01-29-lbrx-asm.ll
test/CodeGen/PowerPC/2007-01-31-InlineAsmAddrMode.ll
test/CodeGen/X86/2006-07-12-InlineAsmQConstraint.ll for PR828
llvm-svn: 36547
2007-04-28 21:12:06 +00:00
de339fa55d
eliminate more redundant constraint type analysis
...
llvm-svn: 36546
2007-04-28 21:03:16 +00:00
b2e55562ed
merge constraint type analysis stuff together.
...
llvm-svn: 36545
2007-04-28 21:01:43 +00:00
d7e3b6c442
Significant refactoring of the inline asm stuff, to support future changes.
...
No functionality change.
llvm-svn: 36544
2007-04-28 20:49:53 +00:00
b18f8f85e9
Implement review feedback. Aliasees can be either GlobalValue's or
...
bitcasts of them.
llvm-svn: 36537
2007-04-28 13:45:00 +00:00
1deacd61f4
memory inputs to an inline asm are required to have an address available.
...
If the operand is not already an indirect operand, spill it to a constant
pool entry or a stack slot.
This fixes PR1356 and CodeGen/X86/2007-04-27-InlineAsm-IntMemInput.ll
llvm-svn: 36536
2007-04-28 06:42:38 +00:00
d102ed0ac6
Fix CodeGen/Generic/2007-04-27-LargeMemObject.ll and
...
CodeGen/Generic/2007-04-27-InlineAsm-X-Dest.ll
llvm-svn: 36534
2007-04-28 06:08:13 +00:00
4df3e8093b
Fix this to match change to InlineAsm class.
...
llvm-svn: 36524
2007-04-28 04:05:59 +00:00
1cbe208cda
Fix incorrect legalization of EHSELECTOR. This fixes
...
CodeGen/Generic/2007-04-14-EHSelectorCrash.ll and PR1326
llvm-svn: 36510
2007-04-27 17:12:52 +00:00
bf535fc8bd
Expand UINT_TO_FP in turns of SINT_TO_FP when UINTTOFP_* libcalls are not available.
...
llvm-svn: 36501
2007-04-27 07:33:31 +00:00
784fe9dbbb
improve EH global handling, patch by Duncan Sands.
...
llvm-svn: 36499
2007-04-27 01:20:11 +00:00
8131ab7c0f
enable Anton's shift/and switch lowering stuff! It now passes ppc bootstrap
...
successfully! woohoo...
llvm-svn: 36496
2007-04-26 21:09:43 +00:00
d7ae7f1659
Fixx off-by-one bug, which prevents llvm-gcc bootstrap on ppc32
...
llvm-svn: 36490
2007-04-26 20:44:04 +00:00
e131e3ac02
Fix a typo in a comment.
...
llvm-svn: 36485
2007-04-26 19:40:56 +00:00
910c80851e
Rename findRegisterUseOperand to findRegisterUseOperandIdx to avoid confusion.
...
llvm-svn: 36483
2007-04-26 19:00:32 +00:00
ed23a1387e
Minor bug.
...
llvm-svn: 36473
2007-04-26 08:24:22 +00:00
43a17fe826
Be careful when to add implicit kill / dead operands. Don't add them during / post reg-allocation.
...
llvm-svn: 36458
2007-04-26 01:40:09 +00:00
0ba174534c
Match MachineFunction::UsedPhysRegs changes.
...
llvm-svn: 36452
2007-04-25 22:13:27 +00:00
d21968d11a
Change UsedPhysRegs from array bool to BitVector to save some space. Setting / getting its states now go through MachineFunction.
...
llvm-svn: 36451
2007-04-25 22:10:09 +00:00
d4549c5527
Clean up.
...
llvm-svn: 36449
2007-04-25 21:34:08 +00:00
0fbe14ab87
Data structure change to improve compile time (especially in debug mode).
...
llvm-svn: 36447
2007-04-25 19:34:00 +00:00
15f269afa3
This was lefted out. Fixed sumarray-dbl.
...
llvm-svn: 36445
2007-04-25 18:33:21 +00:00
a97b694c82
Implement aliases. This fixes PR1017 and it's dependent bugs. CFE part
...
will follow.
llvm-svn: 36435
2007-04-25 14:27:10 +00:00
7818c03c6b
Fix for PR1306.
...
- A register def / use now implicitly affects sub-register liveness but does
not affect liveness information of super-registers.
- Def of a larger register (if followed by a use later) is treated as
read/mod/write of a smaller register.
llvm-svn: 36434
2007-04-25 07:30:23 +00:00
11dc5abde1
Clean up.
...
llvm-svn: 36431
2007-04-25 07:18:20 +00:00
b975bebec1
support for >4G stack frames
...
llvm-svn: 36425
2007-04-25 04:30:24 +00:00
9bd98ea4c1
support > 4G stack objects
...
llvm-svn: 36422
2007-04-25 04:20:54 +00:00
cb0ed0cfbd
allow support for 64-bit stack objects
...
llvm-svn: 36420
2007-04-25 04:08:28 +00:00
01a26c74ae
Be more careful about folding op(x, undef) when we have vector operands.
...
This fixes CodeGen/X86/2007-04-24-VectorCrash.ll
llvm-svn: 36413
2007-04-25 00:00:45 +00:00
47917b697f
Assertion when using a 1-element vector for an add operation. Get the
...
real vector type in this case.
llvm-svn: 36402
2007-04-24 21:13:23 +00:00
4cfa616cee
Use '-1U' where '-1UL' is obvious overkill, eliminating gcc warnings about
...
tests always being true in the process.
llvm-svn: 36387
2007-04-24 01:24:20 +00:00
8653d29b45
modify per review commentary
...
llvm-svn: 36383
2007-04-23 23:33:31 +00:00
14a28f13c8
make EmitAlignment work the way Chris says it should
...
llvm-svn: 36368
2007-04-23 19:58:54 +00:00
8af6d5896f
PR400 phase 2. Propagate attributed load/store information through DAGs.
...
llvm-svn: 36356
2007-04-22 23:15:30 +00:00
4e91908f17
X86 TLS: Implement review feedback.
...
llvm-svn: 36318
2007-04-21 20:56:26 +00:00
0c1349e6bc
Revert Christopher Lamb's load/store alignment changes.
...
llvm-svn: 36309
2007-04-21 18:36:27 +00:00
bff50208c8
add support for alignment attributes on load/store instructions
...
llvm-svn: 36301
2007-04-21 08:16:25 +00:00
94314be0e0
Allow the lowering of ISD::GLOBAL_OFFSET_TABLE.
...
llvm-svn: 36290
2007-04-20 23:02:39 +00:00
2518889872
Implement "general dynamic", "initial exec" and "local exec" TLS models for
...
X86 32 bits.
llvm-svn: 36283
2007-04-20 21:38:10 +00:00
4c53d321aa
VarInfo::UsedBlocks is no longer used. Remove.
...
llvm-svn: 36250
2007-04-18 05:04:38 +00:00
f03c90bee6
allow SRL to simplify its operands, as it doesn't demand all bits as input.
...
llvm-svn: 36245
2007-04-18 03:06:49 +00:00
bf14f20632
When replacing a node in SimplifyDemandedBits, if the old node used any
...
single-use nodes, they will be dead soon. Make sure to remove them before
processing other nodes. This implements CodeGen/X86/shl_elim.ll
llvm-svn: 36244
2007-04-18 03:05:22 +00:00
15c1b820cc
fix a pasto
...
llvm-svn: 36242
2007-04-18 03:01:40 +00:00
b408e8f11d
Don't populate TryAgainList when coalescing only physical registers with virtual registers.
...
llvm-svn: 36240
2007-04-18 02:30:19 +00:00
a5a0c7c909
Increment use count of new virtuals created during PHI elimination.
...
llvm-svn: 36233
2007-04-18 00:36:11 +00:00
4aff52bf3d
Fix a bug in my previous patch, grabbing the shift amount width from the
...
wrong operand.
llvm-svn: 36223
2007-04-17 22:53:02 +00:00
9a861a8550
Fold (x << c1)>> c2 into a single shift if the bits shifted out aren't used.
...
This compiles:
int baz(long long a) { return (short)(((int)(a >>24)) >> 9); }
into:
_baz:
srwi r2, r3, 1
extsh r3, r2
blr
on PPC, instead of:
_baz:
slwi r2, r3, 8
srwi r2, r2, 9
extsh r3, r2
blr
GCC produces:
_baz:
srwi r10,r4,24
insrwi r10,r3,24,0
srawi r9,r3,24
srawi r3,r10,9
extsh r3,r3
blr
This implements CodeGen/PowerPC/shl_elim.ll
llvm-svn: 36221
2007-04-17 21:14:16 +00:00
b881bdabd9
Copy coalescing change to prevent a physical register from being pin to a
...
long live interval that has low usage density.
1. Change order of coalescing to join physical registers with virtual
registers first before virtual register intervals become too long.
2. Check size and usage density to determine if it's worthwhile to join.
3. If joining is aborted, assign virtual register live interval allocation
preference field to the physical register.
4. Register allocator should try to allocate to the preferred register
first (if available) to create identify moves that can be eliminated.
llvm-svn: 36218
2007-04-17 20:32:26 +00:00
57b5214d59
Add a register allocation preference field; add a method to compute size of a live interval.
...
llvm-svn: 36216
2007-04-17 20:25:11 +00:00
8387cf1100
Keep track of number of uses within the function per virtual register.
...
llvm-svn: 36214
2007-04-17 20:22:11 +00:00
9ad5915559
SIGN_EXTEND_INREG does not demand its top bits. Give SimplifyDemandedBits
...
a chance to hack on it. This compiles:
int baz(long long a) { return (short)(((int)(a >>24)) >> 9); }
into:
_baz:
slwi r2, r3, 8
srwi r2, r2, 9
extsh r3, r2
blr
instead of:
_baz:
srwi r2, r4, 24
rlwimi r2, r3, 8, 0, 23
srwi r2, r2, 9
extsh r3, r2
blr
This implements CodeGen/PowerPC/sign_ext_inreg1.ll
llvm-svn: 36212
2007-04-17 19:03:21 +00:00
58ec791b4c
Fix problems in the PartSet lowering having to do with incorrect bit width.
...
llvm-svn: 36180
2007-04-16 22:21:14 +00:00
fb80151c42
Removed tabs everywhere except autogenerated & external files. Add make
...
target for tabs checking.
llvm-svn: 36146
2007-04-16 18:10:23 +00:00
6bd7b7b30b
disable switch lowering using shift/and. It still breaks ppc bootstrap for
...
some reason. :( Will investigate.
llvm-svn: 36011
2007-04-14 19:39:41 +00:00
8a1a84f96e
Fix PR1325: Case range optimization was performed in the case it
...
shouldn't. Also fix some "latent" bug on 64-bit platforms
llvm-svn: 35990
2007-04-14 13:25:55 +00:00
7196f09edc
disable shift/and lowering to work around PR1325 for now.
...
llvm-svn: 35985
2007-04-14 02:26:56 +00:00
e288040abf
Fix PR1323 : we haven't updated phi nodes in good manner :)
...
llvm-svn: 35963
2007-04-13 06:53:51 +00:00
1b9213730f
Make sure intrinsics that are lowered to functions make the function weak
...
linkage so we only end up with one of them in a program. These are, after
all overloaded and templatish in nature.
llvm-svn: 35956
2007-04-12 21:53:38 +00:00
0f2f65f723
Fix bugs in generated code for part_select and part_set so that llc doesn't
...
barf when CBE is run with a program that contains these intrinsics.
llvm-svn: 35946
2007-04-12 13:30:14 +00:00
83faeb7611
Fix a bug in PartSet. The replacement value needs to be zext or trunc to
...
the size of the value, not just zext. Also, give better names to two BBs.
llvm-svn: 35945
2007-04-12 12:46:33 +00:00
5111499136
the result of an inline asm copy can be an arbitrary VT that the register
...
class supports. In the case of vectors, this means we often get the wrong
type (e.g. we get v4f32 instead of v8i16). Make sure to convert the vector
result to the right type. This fixes CodeGen/X86/2007-04-11-InlineAsmVectorResult.ll
llvm-svn: 35944
2007-04-12 06:00:20 +00:00
a77cb3ce68
fold noop vbitconvert instructions
...
llvm-svn: 35943
2007-04-12 05:58:43 +00:00
784a68a702
Fix weirdness handling single element vectors.
...
llvm-svn: 35941
2007-04-12 04:44:28 +00:00
c6251a7dfd
For PR1284:
...
Implement the "part_set" intrinsic.
llvm-svn: 35938
2007-04-12 02:48:46 +00:00
18e4ac4107
fix an infinite loop compiling ldecod, notice by JeffC.
...
llvm-svn: 35910
2007-04-11 16:51:53 +00:00
a083ffcad7
Fix this harder.
...
llvm-svn: 35888
2007-04-11 06:50:51 +00:00
c5f85d3738
don't create shifts by zero, fix some problems with my previous patch
...
llvm-svn: 35887
2007-04-11 06:43:25 +00:00
65786b078c
Teach the codegen to turn [aez]ext (setcc) -> selectcc of 1/0, which often
...
allows other simplifications. For example, this compiles:
int isnegative(unsigned int X) {
return !(X < 2147483648U);
}
Into this code:
x86:
movl 4(%esp), %eax
shrl $31, %eax
ret
arm:
mov r0, r0, lsr #31
bx lr
thumb:
lsr r0, r0, #31
bx lr
instead of:
x86:
cmpl $0, 4(%esp)
sets %al
movzbl %al, %eax
ret
arm:
mov r3, #0
cmp r0, #0
movlt r3, #1
mov r0, r3
bx lr
thumb:
mov r2, #1
mov r1, #0
cmp r0, #0
blt LBB1_2 @entry
LBB1_1: @entry
cpy r2, r1
LBB1_2: @entry
cpy r0, r2
bx lr
Testcase here: test/CodeGen/Generic/ispositive.ll
llvm-svn: 35883
2007-04-11 05:32:27 +00:00
41189c63cc
Codegen integer abs more efficiently using the trick from the PPC CWG. This
...
improves codegen on many architectures. Tests committed as CodeGen/*/iabs.ll
X86 Old: X86 New:
_test: _test:
movl 4(%esp), %ecx movl 4(%esp), %eax
movl %ecx, %eax movl %eax, %ecx
negl %eax sarl $31, %ecx
testl %ecx, %ecx addl %ecx, %eax
cmovns %ecx, %eax xorl %ecx, %eax
ret ret
PPC Old: PPC New:
_test: _test:
cmpwi cr0, r3, -1 srawi r2, r3, 31
neg r2, r3 add r3, r3, r2
bgt cr0, LBB1_2 ; xor r3, r3, r2
LBB1_1: ; blr
mr r3, r2
LBB1_2: ;
blr
ARM Old: ARM New:
_test: _test:
rsb r3, r0, #0 add r3, r0, r0, asr #31
cmp r0, #0 eor r0, r3, r0, asr #31
movge r3, r0 bx lr
mov r0, r3
bx lr
Thumb Old: Thumb New:
_test: _test:
neg r2, r0 asr r2, r0, #31
cmp r0, #0 add r0, r0, r2
bge LBB1_2 eor r0, r2
LBB1_1: @ bx lr
cpy r0, r2
LBB1_2: @
bx lr
Sparc Old: Sparc New:
test: test:
save -96, %o6, %o6 save -96, %o6, %o6
sethi 0, %l0 sra %i0, 31, %l0
sub %l0, %i0, %l0 add %i0, %l0, %l1
subcc %i0, -1, %l1 xor %l1, %l0, %i0
bg .BB1_2 restore %g0, %g0, %g0
nop retl
.BB1_1: nop
or %g0, %l0, %i0
.BB1_2:
restore %g0, %g0, %g0
retl
nop
It also helps alpha/ia64 :)
llvm-svn: 35881
2007-04-11 05:11:38 +00:00
a472f66dd0
For PR1146:
...
Put the parameter attributes in their own ParamAttr name space. Adjust the
rest of llvm as a result.
llvm-svn: 35877
2007-04-11 02:44:20 +00:00
f269d84ca0
apparently some people commit without building the tree, or they forget to
...
commit a LOT of files.
llvm-svn: 35858
2007-04-10 03:20:39 +00:00
e0bbbd3774
No longer needed.
...
llvm-svn: 35850
2007-04-09 23:42:32 +00:00
35f0417ec1
remove dead target hooks.
...
llvm-svn: 35847
2007-04-09 23:34:08 +00:00
39f65335d5
remove some dead target hooks, subsumed by isLegalAddressingMode
...
llvm-svn: 35840
2007-04-09 22:27:04 +00:00
da964a2852
Use integer log for metric calculation
...
llvm-svn: 35834
2007-04-09 21:57:03 +00:00
0475f3b4e9
Unbreak VC++ build.
...
llvm-svn: 35817
2007-04-09 14:32:59 +00:00
506eaf7915
Next stage into switch lowering refactoring
...
1. Fix some bugs in the jump table lowering threshold
2. Implement much better metric for optimal pivot selection
3. Tune thresholds for different lowering methods
4. Implement shift-and trick for lowering small (<machine word
length) cases with few destinations. Good testcase will follow.
llvm-svn: 35816
2007-04-09 12:31:58 +00:00
71b79e3d99
For PR1146:
...
Adapt handling of parameter attributes to use the new ParamAttrsList class.
llvm-svn: 35814
2007-04-09 06:17:21 +00:00
7b2decfa0a
implement CodeGen/X86/inline-asm-x-scalar.ll:test3
...
llvm-svn: 35802
2007-04-09 05:31:20 +00:00
18d6718e78
add some assertions
...
llvm-svn: 35800
2007-04-09 05:23:13 +00:00
f73d215023
Fix a bug introduced with my previous patch, where it didn't correctly handle
...
instructions which replace themselves when FI's are rewritten (common on ppc).
This fixes CodeGen/PowerPC/2006-10-17-ppc64-alloca.ll
llvm-svn: 35789
2007-04-09 01:19:33 +00:00
0df5357436
Fix CodeGen/Generic/2007-04-08-MultipleFrameIndices.ll and PR1308:
...
some instructions can have multiple frame indices in them. If this happens,
rewrite all of them.
llvm-svn: 35785
2007-04-09 00:46:10 +00:00
b49917da92
Fix PR1316
...
llvm-svn: 35783
2007-04-09 00:33:58 +00:00
e55ecfb870
Fix for CodeGen/X86/2007-04-08-InlineAsmCrash.ll and PR1314
...
llvm-svn: 35779
2007-04-08 22:23:26 +00:00
1c741e95d3
minor comment fix
...
llvm-svn: 35696
2007-04-06 17:47:14 +00:00
85460acfbf
Change the bit_part_select (non)implementation from "return 0" to abort.
...
llvm-svn: 35679
2007-04-05 01:20:18 +00:00
cce90f55ed
Implement the llvm.bit.part_select.iN.iN.iN overloaded intrinsic.
...
llvm-svn: 35678
2007-04-04 23:48:25 +00:00
915e61736b
Properly emit range comparisons for switch cases, where neighbour cases
...
go to the same destination. Now we're producing really good code for
switch-lower-feature.ll testcase
llvm-svn: 35672
2007-04-04 21:14:49 +00:00
8be98c1572
Re-materialize all loads from fixed stack slots.
...
llvm-svn: 35660
2007-04-04 07:40:01 +00:00
2bf2aadd9a
Trivially re-materializable instructions have spill weights that are half of what it would be otherwise.
...
llvm-svn: 35658
2007-04-04 07:04:55 +00:00
06a7041ff9
Bad bad bug. findRegisterUseOperand() returns -1 if a use if not found.
...
llvm-svn: 35618
2007-04-03 06:43:29 +00:00
16627a542f
1. Insert custom lowering hooks for ISD::ROTR and ISD::ROTL.
...
2. Help DAGCombiner recognize zero/sign/any-extended versions of ROTR and ROTL
patterns. This was motivated by the X86/rotate.ll testcase, which should now
generate code for other platforms (and soon-to-come platforms.) Rewrote code
slightly to make it easier to read.
llvm-svn: 35605
2007-04-02 21:36:32 +00:00
476fb6a5c9
Ugh. Copy coalescer does not update register numbers.
...
llvm-svn: 35600
2007-04-02 18:49:18 +00:00
fad9bd6b92
For PR1297:
...
Make sure that the CTPOP result is casted to i32 as the bit counting
intrinsics all return i32 now (this affects CTLZ and CTTZ as well).
llvm-svn: 35567
2007-04-02 01:01:49 +00:00
6bba6c8143
For PR1297:
...
Support overloaded intrinsics bswap, ctpop, cttz, ctlz.
llvm-svn: 35547
2007-04-01 07:35:23 +00:00
3a0843e734
For PR1297:
...
Adjust for changes in the bit counting intrinsics. They all return i32
now so we have to trunc/zext the DAG node accordingly.
llvm-svn: 35546
2007-04-01 07:34:11 +00:00
a090ffb2ab
For PR1297:
...
Change getOperationName to return std::string instead of const char*
llvm-svn: 35545
2007-04-01 07:32:19 +00:00
f6a6d3c8b0
move a bunch of code out of the sdisel pass into its own opt pass "codegenprepare".
...
llvm-svn: 35529
2007-03-31 04:18:03 +00:00
f2d71d49e2
switch TL::getValueType to use MVT::getValueType.
...
llvm-svn: 35527
2007-03-31 04:05:24 +00:00
bafc837c83
Add a -print-lsr-output option to LLC, to print the output of the LSR pass.
...
llvm-svn: 35522
2007-03-31 00:24:43 +00:00
ac3f81508c
add one addressing mode description hook to rule them all.
...
llvm-svn: 35520
2007-03-30 23:14:50 +00:00
4bbd2eefba
Fix incorrect combination of different loads. Reenable zext-over-truncate
...
combination.
llvm-svn: 35517
2007-03-30 21:38:07 +00:00
9a2a7b174a
Don't add the same MI to register reuse "last def/use" twice if it reads the
...
register more than once.
llvm-svn: 35513
2007-03-30 20:21:35 +00:00
48dd5e7d44
Bug fix for PR1279. When isDead is propagate by copy coalescing, we keep length
...
of dead def live interval at 1 to avoid multiple def's targeting the same
register. The previous patch missed a case where the source operand is live-in.
In that case, remove the whole interval.
llvm-svn: 35512
2007-03-30 20:18:35 +00:00
ccee35fd0d
Disable load width reduction xform of variant (zext (truncate load x)) for
...
big endian targets until llvm-gcc build issue has been resolved.
llvm-svn: 35449
2007-03-29 07:56:46 +00:00
86de3a009d
New entries.
...
llvm-svn: 35445
2007-03-29 02:48:56 +00:00
d771485d89
Notes on re-materialization.
...
llvm-svn: 35420
2007-03-28 08:30:04 +00:00
603e49c6da
Move rematerialization out of beta.
...
llvm-svn: 35419
2007-03-28 08:26:40 +00:00
4388043b25
Scale 1 is always ok.
...
llvm-svn: 35407
2007-03-28 01:55:52 +00:00
c2cba18f2b
Remove isLegalAddressImmediate.
...
llvm-svn: 35406
2007-03-28 01:53:55 +00:00
07c42d43a2
GEP index sinking fixes:
...
1) Take address scale into consideration. e.g. i32* -> scale 4.
2) Examine all the users of GEP.
3) Generalize to inter-block GEP's (no longer uses loopinfo).
4) Don't do xform if GEP has other variable index(es).
llvm-svn: 35403
2007-03-28 01:49:39 +00:00
b41697c006
Fix for PR1279. Dead def has a live interval of length 1. Copy coalescing should
...
not violate that.
llvm-svn: 35396
2007-03-28 01:30:37 +00:00
37a0bfe128
Remove dead code
...
llvm-svn: 35380
2007-03-27 12:05:48 +00:00
3a9d68181a
Split big monster into small helpers. No functionality change.
...
llvm-svn: 35379
2007-03-27 11:29:11 +00:00
c42406b5ad
SDISel does not preserve all, it changes CFG and other info.
...
llvm-svn: 35376
2007-03-27 00:53:36 +00:00
fdbdf43632
Don't call getOperandConstraint() if operand index is greater than
...
TID->numOperands.
llvm-svn: 35375
2007-03-27 00:48:28 +00:00
4a09b1b5be
Fix for PR1266. Don't mark a two address operand IsKill.
...
llvm-svn: 35365
2007-03-26 22:40:42 +00:00
ec3ac316e2
Change findRegisterUseOperand() to return operand index instead.
...
llvm-svn: 35363
2007-03-26 22:37:45 +00:00
fd18a0cdf4
Fix reversed logic in getRegsUsed. Rename RegStates to RegsAvailable to
...
hopefully forestall similar errors.
llvm-svn: 35362
2007-03-26 22:23:54 +00:00
8275f0e0af
SIGN_EXTEND_INREG requires one extra operand, a ValueType node.
...
llvm-svn: 35350
2007-03-26 07:12:51 +00:00
7037826c86
First step of switch lowering refactoring: perform worklist-driven
...
strategy, emit JT's where possible.
llvm-svn: 35338
2007-03-25 15:07:15 +00:00
77f0479833
Implement support for vector operands to inline asm, implementing
...
CodeGen/X86/2007-03-24-InlineAsmVectorOp.ll
llvm-svn: 35332
2007-03-25 05:00:54 +00:00
3d7efa2586
implement initial support for the silly X constraint. Testcase here: CodeGen/X86/2007-03-24-InlineAsmXConstraint.ll
...
llvm-svn: 35327
2007-03-25 04:35:41 +00:00
843e44503c
Implement CodeGen/X86/2007-03-24-InlineAsmMultiRegConstraint.ll
...
llvm-svn: 35324
2007-03-25 02:18:14 +00:00
d685514e2e
switch TargetLowering::getConstraintType to take the entire constraint,
...
not just the first letter. No functionality change.
llvm-svn: 35322
2007-03-25 02:14:49 +00:00
2a991268f7
don't rely on ADL
...
llvm-svn: 35299
2007-03-24 17:37:03 +00:00
b7051f596a
Adjust offset to compensate for big endian machines.
...
llvm-svn: 35293
2007-03-24 00:02:43 +00:00
a883b58caf
Make sure SEXTLOAD of the specific type is supported on the target.
...
llvm-svn: 35289
2007-03-23 22:13:36 +00:00
e2f5f24e8e
Also replace uses of SRL if that's also folded during ReduceLoadWidth().
...
llvm-svn: 35286
2007-03-23 20:55:21 +00:00
a824e79f06
A couple of bug fixes for reducing load width xform:
...
1. Address offset is in bytes.
2. Make sure truncate node uses are replaced with new load.
llvm-svn: 35274
2007-03-23 02:16:52 +00:00
dcb291faa4
Change uses of Function::front to Function::getEntryBlock for readability.
...
llvm-svn: 35265
2007-03-22 16:38:57 +00:00
464dc9b74c
More opportunities to reduce load size.
...
llvm-svn: 35254
2007-03-22 01:54:19 +00:00
a20815e157
Fix for PR1257. Bug in live range shortening as a result of copy coalescing
...
where the destination is dead.
llvm-svn: 35252
2007-03-22 01:26:05 +00:00
0c6bb5eab7
repair x86 performance, dejagnu problems from previous change
...
llvm-svn: 35245
2007-03-21 21:51:52 +00:00
d63baead9b
fold (truncate (srl (load x), c)) -> (smaller load (x+c/vt bits))
...
llvm-svn: 35239
2007-03-21 20:14:05 +00:00
3578dd61c6
Potential spiller improvement.
...
llvm-svn: 35228
2007-03-20 22:22:38 +00:00
bacf4acf65
do not share old induction variables when this would result in invalid
...
instructions (that would have to be split later)
llvm-svn: 35227
2007-03-20 21:54:54 +00:00
d05a1a2ade
maintain LiveIn when splitting blocks (register scavenging needs it)
...
llvm-svn: 35226
2007-03-20 21:35:06 +00:00
1baf5c84ab
Fix some VC++ warnings.
...
llvm-svn: 35224
2007-03-20 20:43:18 +00:00
971aa18867
Code clean up.
...
llvm-svn: 35220
2007-03-20 20:09:03 +00:00
550cf0369c
Minor bug.
...
llvm-svn: 35219
2007-03-20 19:32:11 +00:00
25878b45f5
CopyToReg source operand can be a physical register.
...
llvm-svn: 35213
2007-03-20 16:46:44 +00:00
0e3278e505
First cut trivial re-materialization support.
...
llvm-svn: 35208
2007-03-20 08:13:50 +00:00
25d00d545d
Remove -reduce-joining-phys-regs options. Make it on by default.
...
llvm-svn: 35165
2007-03-19 18:08:26 +00:00
c5e74f6404
Minor bug fix.
...
llvm-svn: 35153
2007-03-19 04:22:35 +00:00
d8f2e4fe4c
- Merge UsedBlocks info after two virtual registers are coalesced.
...
- Use distance to closest use to determine whether to abort coalescing.
llvm-svn: 35141
2007-03-18 09:05:55 +00:00
5382426577
Keep UsedBlocks info accurate.
...
llvm-svn: 35140
2007-03-18 09:02:31 +00:00
f6f043332f
Track the BB's where each virtual register is used.
...
llvm-svn: 35135
2007-03-17 09:29:54 +00:00
7b2a001669
Joining a live interval of a physical register with a virtual one can turn out
...
to be really bad. Once they are joined they are not broken apart. Also, physical
intervals cannot be spilled!
Added a heuristic as a workaround for this. Be careful coalescing with a
physical register if the virtual register uses are "far". Check if there are
uses in the same loop as the source (copy instruction). Check if it is in the
loop preheader, etc.
llvm-svn: 35134
2007-03-17 09:27:35 +00:00
a2465dfc07
Use SmallSet instead of std::set.
...
llvm-svn: 35133
2007-03-17 08:53:30 +00:00
be22235790
If sdisel has decided to sink GEP index expression into any BB. Replace all uses
...
in that BB.
llvm-svn: 35132
2007-03-17 08:22:49 +00:00
c5bc763f50
Turn on GEP index sinking by default.
...
llvm-svn: 35127
2007-03-16 18:32:30 +00:00
0a9d0cabaf
Stupid bug.
...
llvm-svn: 35126
2007-03-16 17:50:20 +00:00
009ea54262
Sink a binary expression into its use blocks if it is a loop invariant
...
computation used as GEP indexes and if the expression can be folded into
target addressing mode of GEP load / store use types.
llvm-svn: 35123
2007-03-16 08:46:27 +00:00
a2a2fd1e55
Added isLegalAddressExpression hook to test if the given expression can be
...
folded into target addressing mode for the given type.
llvm-svn: 35121
2007-03-16 08:42:32 +00:00
88de94a4fb
Debugging output stuff.
...
llvm-svn: 35117
2007-03-15 21:19:28 +00:00
b9e3db67fb
Estimate a cost using the possible number of scratch registers required and use
...
it as a late BURR scheduling tie-breaker.
Intuitively, it's good to push down instructions whose results are liveout so
their long live ranges won't conflict with other values which are needed inside
the BB. Further prioritize liveout instructions by the number of operands which
are calculated within the BB.
llvm-svn: 35109
2007-03-14 22:43:40 +00:00
2874855302
Try schedule def + use closer whne Sethi-Ullman numbers are the same.
...
e.g.
t1 = op t2, c1
t3 = op t4, c2
and the following instructions are both ready.
t2 = op c3
t4 = op c4
Then schedule t2 = op first.
i.e.
t4 = op c4
t2 = op c3
t1 = op t2, c1
t3 = op t4, c2
This creates more short live intervals which work better with the register
allocator.
llvm-svn: 35089
2007-03-13 23:25:11 +00:00
b7004fd889
More flexible TargetLowering LSR hooks for testing whether an immediate is a legal target address immediate or scale.
...
llvm-svn: 35076
2007-03-12 23:37:10 +00:00
df7949a8d0
If a virtual register is already marked alive in this block, that means it is
...
alive in one of the successor block. Do not add it to the kill list.
llvm-svn: 35041
2007-03-09 09:48:56 +00:00
Evan Cheng