RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2025-01-09 21:32:49 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
39,725 Commits 4 Branches 5 Tags 634 MiB
06545d1c4a
Commit Graph

23631 Commits

Author SHA1 Message Date
Duncan Sands
e7f2c80a8c Port some integer multiplication fixes from LegalizeDAG. 
Bail out with an error if there is no libcall available
for the given size of integer.

llvm-svn: 52622
 2008-06-23 15:15:44 +00:00
Duncan Sands
86e6f1a3b6 Support for expanding the result of EXTRACT_ELEMENT. 
llvm-svn: 52621
 2008-06-23 15:08:15 +00:00
Duncan Sands
1b83cbdebe Cleanup up LegalizeTypes handling of loads and 
stores.

llvm-svn: 52620
 2008-06-23 14:19:45 +00:00
Owen Anderson
7d4e8f976c At Chris' suggestion, move the liveness and worklist datastructures into 
instance variables so they can be allocated just once, and reuse the worklist
as the dead list as well.

llvm-svn: 52618
 2008-06-23 06:13:12 +00:00
Dan Gohman
62d8bc0480 Improve LSR's dead-phi detection to handle use-def cycles 
with more than two nodes.

llvm-svn: 52617
 2008-06-22 20:44:02 +00:00
Dan Gohman
4ada89e819 Use Loop::block_iterator. 
llvm-svn: 52616
 2008-06-22 20:18:58 +00:00
Dan Gohman
90894ac18b Generalize createSCEV to be able to form SCEV expressions from 
ConstantExprs.

llvm-svn: 52615
 2008-06-22 19:56:46 +00:00
Dan Gohman
81c83d9a1d Use SCEVAddRecExpr::isAffine. 
llvm-svn: 52614
 2008-06-22 19:23:09 +00:00
Dan Gohman
c1aa753f00 Remove unnecessary #includes. 
llvm-svn: 52613
 2008-06-22 19:21:26 +00:00
Dan Gohman
311865bd24 Move a few more SCEVExpander methods out-of-line. 
llvm-svn: 52612
 2008-06-22 19:09:18 +00:00
Chris Lattner
d80c865a09 Fix PR2369 by making scalarrepl more careful about promoting 
structures.  Its default threshold is to promote things that are
smaller than 128 bytes, which is sane.  However, it is not sane
to do this for things that turn into 128 *registers*.  Add a cap
on the number of registers introduced, defaulting to 128/4=32.

llvm-svn: 52611
 2008-06-22 17:46:21 +00:00
Duncan Sands
d1d3e67d30 Make custom lowering of ADD work correctly. This 
fixes PR2476; patch by Richard Osborne.  The same
problem exists for a bunch of other operators, but
I'm ignoring this because they will be automagically
fixed when the new LegalizeTypes infrastructure lands,
since it already solves this problem centrally.

llvm-svn: 52610
 2008-06-22 09:42:16 +00:00
Eli Friedman
369401ef95 Fix for PR2479: correctly optimize expressions like (a > 13) & (a == 
15).

See also PR1800, which is about the signed case.

llvm-svn: 52608
 2008-06-21 23:36:13 +00:00
Dan Gohman
91a43b082c Use Instruction::eraseFromParent(). 
llvm-svn: 52606
 2008-06-21 22:08:46 +00:00
Dan Gohman
e0744f181b Use Function's arg_size() and size() methods. 
llvm-svn: 52605
 2008-06-21 22:06:54 +00:00
Dan Gohman
a41cf16a8f Simplify some getNode calls. 
llvm-svn: 52604
 2008-06-21 22:06:07 +00:00
Dan Gohman
4d2c1733d7 canClobberPhysRegDefs shouldn't called without checking hasPhysRegDefs; 
check this with an assert.

llvm-svn: 52603
 2008-06-21 22:05:24 +00:00
Dan Gohman
e6da01721b Avoid creating a redundant zero APInt. 
llvm-svn: 52602
 2008-06-21 22:03:12 +00:00
Dan Gohman
a313808e69 Use clear() to zero an existing APInt. 
llvm-svn: 52601
 2008-06-21 22:02:15 +00:00
Dan Gohman
7460588026 Use back() instead of [size()-1]. 
llvm-svn: 52600
 2008-06-21 22:00:54 +00:00
Dan Gohman
bed65a69c3 Use MachineBasicBlock::transferSuccessors. 
llvm-svn: 52594
 2008-06-21 20:21:19 +00:00
Dan Gohman
9b9cf4cf9c Use static_cast instead of reinterpret_cast for casting void*. 
llvm-svn: 52592
 2008-06-21 20:17:03 +00:00
Chris Lattner
5b0bc729b3 Fix warning when assertions disabled. 
llvm-svn: 52590
 2008-06-21 19:49:01 +00:00
Chris Lattner
76d8bbc7c2 fix warning when assertions disabled. 
llvm-svn: 52589
 2008-06-21 19:48:22 +00:00
Chris Lattner
d3308e835f fix warning when assertion disabled. 
llvm-svn: 52588
 2008-06-21 19:47:44 +00:00
Chris Lattner
96b3955e9b fix some warnings when assertions are disabled. 
llvm-svn: 52587
 2008-06-21 19:47:03 +00:00
Dan Gohman
1e71795b8d Remove a redundant return. 
llvm-svn: 52585
 2008-06-21 19:34:57 +00:00
Dan Gohman
015a5ca374 Remove ScheduleDAG's SUnitMap altogether. Instead, use SDNode's NodeId 
field, which is otherwise unused after instruction selection, as an index
into the SUnit array.

llvm-svn: 52583
 2008-06-21 19:18:17 +00:00
Dan Gohman
4099d0bf27 Add a priority queue class, which is a wrapper around std::priority_queue 
and provides fairly efficient removal of arbitrary elements. Switch
ScheduleDAGRRList from std::set to this new priority queue.

llvm-svn: 52582
 2008-06-21 18:35:25 +00:00
Duncan Sands
1dd6ef8f8e Support for load/store of expanded float types. I 
don't know if a truncating store is possible here,
but added support for it anyway.

llvm-svn: 52577
 2008-06-21 17:00:47 +00:00
Dan Gohman
3a5a96e045 Change ScheduleDAG's SUnitMap from DenseMap<SDNode*, vector<SUnit*> > 
to DenseMap<SDNode*, SUnit*>, and adjust the way cloned SUnit nodes are
handled so that only the original node needs to be in the map.
This speeds up llc on 447.dealII.llvm.bc by about 2%.

llvm-svn: 52576
 2008-06-21 15:52:51 +00:00
Evan Cheng
2ff47edc37 Enable PRE. 
llvm-svn: 52574
 2008-06-21 07:26:53 +00:00
Evan Cheng
1d07cd32c2 Undo spill weight tweak. Need to investigate the performance regressions. 
llvm-svn: 52572
 2008-06-21 06:45:54 +00:00
Dan Gohman
62fb667283 Simplify some template parameterization. 
llvm-svn: 52571
 2008-06-21 01:08:22 +00:00
Evan Cheng
b65bceda9c Back out Matthijs' DAE patches. It's miscompiling gcc driver. 
llvm-svn: 52570
 2008-06-21 00:31:44 +00:00
Evan Cheng
c1f551b43e Enhanced heuristic to determine the *best* register to spill. Instead of picking the register with the lowest spill weight. Consider (up to) 2 additional registers with spill weights that are close to the lowest spill weight. The one with fewest defs and uses that conflicts with the current interval (weighted by loop depth) is the spill candidate. 
This is not always a win, but there are much more wins than loses and wins tend to be more noticeable.

llvm-svn: 52554
 2008-06-20 21:45:16 +00:00
Duncan Sands
59b9db1fb6 Share some code that is common between integer and 
float expansion (and sometimes vector splitting too).

llvm-svn: 52548
 2008-06-20 18:40:50 +00:00
Duncan Sands
bccd4c58ba Rename the operation of turning a float type into an 
integer of the same type.  Before it was "promotion",
but this is confusing because it is quite different
to promotion of integers.  Call it "softening" instead,
inspired by "soft float".

llvm-svn: 52546
 2008-06-20 17:49:55 +00:00
Dan Gohman
10e3188458 Clean up some uses of std::distance, now that we have allnodes_size. 
llvm-svn: 52545
 2008-06-20 17:15:19 +00:00
Dan Gohman
6a75f18b3b Clean up a use of std::distance. 
llvm-svn: 52544
 2008-06-20 17:11:32 +00:00
Dan Gohman
fcc6f11fc3 Tidy up some commments and use the getAggregateOperand and 
getInsertedValueOperand accessors. Thanks Matthijs!

llvm-svn: 52543
 2008-06-20 16:41:17 +00:00
Dan Gohman
3968f90c4c Fix the conditions under which SCCP should examine insertvalue 
instructions. Thanks to Matthijs Kooijman for pointing this out!

llvm-svn: 52542
 2008-06-20 16:39:44 +00:00
Matthijs Kooijman
54efc9f60e 80 column and trailing whitespace fixes. 
llvm-svn: 52539
 2008-06-20 15:34:07 +00:00
Matthijs Kooijman
7a75a306ff Don't let DeadArgumentElimination attempt to update callers when the return 
type wasn't changed.

llvm-svn: 52538
 2008-06-20 15:25:43 +00:00
Matthijs Kooijman
9a9a6f1122 Don't let DeadArgElimination change the return type ({} into void and {T} 
into T) when no return values are actually dead.

llvm-svn: 52537
 2008-06-20 15:16:45 +00:00
Matthijs Kooijman
1d019c28b3 Explicitely track if any arguments or return values were removed in 
DeadArgumentElimination and assert that the function type does not change if
nothing was changed. This should catch subtle changes in function type that are
not intended.

llvm-svn: 52536
 2008-06-20 14:28:52 +00:00
Matthijs Kooijman
35ac78c162 Remove debug output. 
llvm-svn: 52535
 2008-06-20 14:03:35 +00:00
Matthijs Kooijman
a3222e3730 Recommit r52459, rewriting of the dead argument elimination pass. 
This is a fixed version that no longer uses multimap::equal_range, which
resulted in a pointer invalidation problem.

Also, DAE::InspectedFunctions was not really necessary, so it got removed.

Lastly, this version no longer applies the extra arg hack on functions who did
not have any arguments to start with.

llvm-svn: 52532
 2008-06-20 09:36:16 +00:00
Owen Anderson
5213d215e0 Really disable PRE. 
llvm-svn: 52531
 2008-06-20 08:59:13 +00:00
Chris Lattner
9ab7735924 Fix an error handling redefinition of linkonce functions where the 
types differ.  Patch by Nathan Keynes!

llvm-svn: 52527
 2008-06-20 05:29:39 +00:00
Chris Lattner
e588f546c5 Fix PR2471, which is a bug involving an invalid promotion from a conditional load. 
llvm-svn: 52525
 2008-06-20 05:12:56 +00:00
Gordon Henriksen
50ca4ebe97 Add C binding for ExecutionEngine::addGlobalMapping. 
llvm-svn: 52523
 2008-06-20 02:16:11 +00:00
Dan Gohman
359de5a04e Teach ReturnInst lowering about aggregate return values. 
llvm-svn: 52522
 2008-06-20 01:29:26 +00:00
Owen Anderson
3231e506cb Change around the data structures used to store availability sets, resulting in a GVN+PRE that is faster that GVN alone was before. 
llvm-svn: 52521
 2008-06-20 01:15:47 +00:00
Dan Gohman
b4ee9dfc2d Teach SCCP about insertvalue and extractvalue, and about 
propagating constants across aggregate return values when
insertvalue and extractvalue are used.

llvm-svn: 52520
 2008-06-20 01:15:44 +00:00
Dan Gohman
029d1e045e Teach InlineFunction how to differentiate between multiple-value 
return statements and aggregate returns so that it handles both
correctly.

llvm-svn: 52519
 2008-06-20 01:03:44 +00:00
Evan Cheng
d4c055e1f8 Disable PRE for now. It seems to be breaking llvm-gcc bootstrapping. 
llvm-svn: 52518
 2008-06-20 01:01:07 +00:00
Dan Gohman
0d86d31373 Fix the index calculations for the extractvalue lowering code. 
llvm-svn: 52517
 2008-06-20 00:54:19 +00:00
Dan Gohman
826ad19523 Simplify the ComputeLinearIndex logic and fix a few bugs. 
llvm-svn: 52516
 2008-06-20 00:53:00 +00:00
Dan Gohman
bcddb92f4e Simplify this code. Thanks Chris! 
llvm-svn: 52514
 2008-06-20 00:47:44 +00:00
Evan Cheng
4006f4cdf0 ISD::UNDEF should be expanded recursively / iteratively. 
llvm-svn: 52508
 2008-06-19 22:01:11 +00:00
Owen Anderson
c979fcf8ff Add a hidden -disable-pre flag for testing purposes. This should be removed 
once benchmarking is completed.

llvm-svn: 52506
 2008-06-19 19:57:25 +00:00
Owen Anderson
35fb8dbfef PRE requires that critical edges be split. 
llvm-svn: 52505
 2008-06-19 19:54:19 +00:00
Bill Wendling
68bdf9f6d4 Remove dead code causing a warning. 
llvm-svn: 52502
 2008-06-19 18:00:44 +00:00
Dan Gohman
98d028eb42 Use the common API for adding instructions to basic blocks instead of 
using BasicBlock::getInstList.

llvm-svn: 52500
 2008-06-19 17:53:32 +00:00
Owen Anderson
87d761460e Be sure to remove values from the value numbering table after we delete them. 
This fixes a failure on povray.

llvm-svn: 52499
 2008-06-19 17:53:26 +00:00
Dan Gohman
a18aa3f3a2 Use Instruction::moveBefore instead of manipulating the instruction list 
directly.

llvm-svn: 52498
 2008-06-19 17:47:47 +00:00
Dan Gohman
01679b25c0 Avoid using BasicBlock::getInstList directly in a few places. 
llvm-svn: 52497
 2008-06-19 17:37:25 +00:00
Owen Anderson
aa850d32db Revert support for insertvalue and extractvalue instructions for the moment. 
GVN expects that all inputs which to an instruction fall somewhere in the value
hierarchy, which isn't true for these.

llvm-svn: 52496
 2008-06-19 17:25:39 +00:00
Dan Gohman
cd78a0e385 Use the transferSuccessors helper function. 
llvm-svn: 52495
 2008-06-19 17:22:29 +00:00
Dan Gohman
68c5bdd4f5 Delete dead code. 
llvm-svn: 52494
 2008-06-19 17:18:39 +00:00
Matthijs Kooijman
41c5b9a77c Use a CallSite to find the nth argument of a call/invoke instruction instead of 
using getOperand() directly. This makes things work with invoke instructions as
well.

llvm-svn: 52489
 2008-06-19 08:53:24 +00:00
Duncan Sands
054b82d555 Fix some warnings reported by gcc-4.3. Hopefully 
this still compiles on windows - I can't test!

llvm-svn: 52488
 2008-06-19 08:47:31 +00:00
Evan Cheng
b12290e31b Missed a check. 
llvm-svn: 52487
 2008-06-19 06:17:19 +00:00
Eli Friedman
570aa6f801 Fix a bug with <8 x i16> shuffle lowering on X86 where parts of the 
shuffle could be skipped.  The check is invalid because the loop index i 
doesn't correspond to the element actually inserted. The correct check is
already done a few lines earlier, for whether the element is already in 
the right spot, so this shouldn't have any effect on the codegen for 
code that was already correct.

llvm-svn: 52486
 2008-06-19 06:09:51 +00:00
Owen Anderson
864a2c23ed Revert my last patch, which was causing regression test failures. 
llvm-svn: 52485
 2008-06-19 05:29:34 +00:00
Evan Cheng
22b431d28f Coalesce copy from one register class to a sub register class. e.g. X86::MOV16to16_. 
llvm-svn: 52480
 2008-06-19 01:39:21 +00:00
Evan Cheng
f129d15d03 Cosmetic changes. 
llvm-svn: 52479
 2008-06-19 01:21:26 +00:00
Evan Cheng
4416f16a6a Unneeded include's. 
llvm-svn: 52478
 2008-06-19 01:21:02 +00:00
Evan Cheng
968679913e Minor spiller tweak to unfavor reload into load/store instructions. 
llvm-svn: 52477
 2008-06-19 01:16:17 +00:00
Owen Anderson
abbc43e493 Insert empty slots into the instruction numbering in live intervals, so that we can more easily 
add new instructions.

llvm-svn: 52475
 2008-06-19 00:10:49 +00:00
Owen Anderson
5b8d39569b Add support for extractvalue and insertvalue instructions in GVN. 
llvm-svn: 52472
 2008-06-18 21:59:00 +00:00
Owen Anderson
3f78e260c1 Add local PRE to GVN. This only operates in cases where it would not increase code size, namely when the instantiated expression 
would only need to be created in one predecessor.

llvm-svn: 52471
 2008-06-18 21:41:49 +00:00
Argyrios Kyrtzidis
5ef6b9b977 Fix the source line debug information for the Windows platform. 
According to DWARF-2 specification, the line information is provided through an offset in the .debug_line section.
Replace the label reference that is used with a section offset.

llvm-svn: 52468
 2008-06-18 19:27:37 +00:00
Chris Lattner
f9d9f0ec4c Fix the regressions on sext-misc.ll my patch yesterday caused. 
llvm-svn: 52466
 2008-06-18 18:11:55 +00:00
Owen Anderson
e680ac8dcb Revert r52459, which was causing an infinite loop or massive slowdown on MultiSource/Applications/SPASS, and possibly others as well. 
Please reapply once this is fixed.

llvm-svn: 52465
 2008-06-18 17:32:16 +00:00
Dan Gohman
606a60ad68 Move SCEVExpander::visitAddExpr out-of-line. 
llvm-svn: 52464
 2008-06-18 16:37:11 +00:00
Dan Gohman
82f9df016e Move LSR's private isZero function to a public SCEV member 
function, and make use of it in several places.

llvm-svn: 52463
 2008-06-18 16:23:07 +00:00
Matthijs Kooijman
b7e2818227 Rewrite the DeadArgumentElimination pass, to use a more explicit tracking of 
dependencies between return values and/or arguments. Also make the handling of
arguments and return values the same.

The pass now looks properly inside returned structs, but only at the first                                                                                     
level (ie, not inside nested structs).

Also add a testcase for testing various variations of (multiple) dead rerturn
values.

llvm-svn: 52459
 2008-06-18 11:12:53 +00:00
Matthijs Kooijman
8177832230 Reapply r52397 (make IPConstProp promote returned arguments), but fixed this 
time. Sorry for the trouble!

This time, also add a testcase, which I should have done in the first place...

llvm-svn: 52455
 2008-06-18 08:30:37 +00:00
Evan Cheng
0570953e28 XOR32rr, etc. are not AsCheapAsMove, but MOV32ri, etc. are. 
llvm-svn: 52454
 2008-06-18 08:13:07 +00:00
Matthijs Kooijman
bb59138fa6 Reapply r52396, it was unrelated to the breakage (that was caused by r52397, my 
commit after this).

llvm-svn: 52453
 2008-06-18 08:09:27 +00:00
Evan Cheng
ad0a4a31f8 Complete support for two-address pass rematerialization. Now *almost* always a win. 
llvm-svn: 52452
 2008-06-18 07:49:14 +00:00
Evan Cheng
cba32609d2 Cosmetic. 
llvm-svn: 52450
 2008-06-18 07:47:28 +00:00
Chris Lattner
93da79f7a1 implement some simple bswap optimizations, rdar://5992453 
llvm-svn: 52442
 2008-06-18 04:33:20 +00:00
Chris Lattner
7e403da191 make truncate/sext elimination capable of changing phi's. This 
implements rdar://6013816 and the testcase in Transforms/InstCombine/sext-misc.ll.

llvm-svn: 52440
 2008-06-18 04:00:49 +00:00
Evan Cheng
deb754898b Unbreak DECLARE isel in pic mode. 
llvm-svn: 52439
 2008-06-18 02:48:27 +00:00
Devang Patel
8f157a3670 Preserve dominance frontier while trivially unswitching loop. 
llvm-svn: 52438
 2008-06-18 02:16:38 +00:00
Dan Gohman
c190d65454 Auto-upgrade code for multiple-value return statements. This code 
isn't actually called yet.

llvm-svn: 52435
 2008-06-17 23:38:43 +00:00
Dan Gohman
d9d529d7ed In InsertValueInst's copy ctor, actually copy the operands. 
llvm-svn: 52434
 2008-06-17 23:25:49 +00:00
Owen Anderson
c1ac0c1c41 We don't want to find dependencies within the same block in this case. It leads to incorrect results because 
we're detecting something at or after the call we're querying on.

llvm-svn: 52433
 2008-06-17 22:27:06 +00:00
Dan Gohman
24d9c8400f Implement the ExtractValueInst::getIndexedType that accepts one 
index value.

llvm-svn: 52432
 2008-06-17 21:07:55 +00:00
Evan Cheng
10ead77866 Live-through live interval is [mbb start, mbb end+1]. 
llvm-svn: 52431
 2008-06-17 20:13:36 +00:00
Evan Cheng
3dc2cfe366 When extending a liveinterval by commuting, don't throw away the live ranges that are not affected. 
llvm-svn: 52430
 2008-06-17 20:11:16 +00:00
Ted Kremenek
a9f1f1db52 Consistently set "Buckets[NumBuckets] = reinterpret_cast<void*>(-1)" throughout FoldingSet.cpp. 
llvm-svn: 52425
 2008-06-17 19:12:43 +00:00
Owen Anderson
79605db969 Add an insertBefore method for attaching previously unattached instructions, 
such as those created by clone(), to a basic block.

llvm-svn: 52424
 2008-06-17 18:29:27 +00:00
Evan Cheng
acd28b95da It's not safe to remove SUBREG_TO_REG that looks like identity copies, e.g. movl %eax, %eax on x86-64 actually does a zero-extend. 
llvm-svn: 52421
 2008-06-17 17:59:16 +00:00
Anton Korobeynikov
6a779ad6fa Add one more 'magic' define :) 
llvm-svn: 52420
 2008-06-17 17:57:43 +00:00
Anton Korobeynikov
cdfbb1eae5 Unbreak non-PPC builds 
llvm-svn: 52419
 2008-06-17 17:38:31 +00:00
Anton Korobeynikov
5b04f6d658 Provide generic hooks for icache invalidation. Add PPC implementation. 
Patch by Gary Benson!

llvm-svn: 52418
 2008-06-17 17:30:05 +00:00
Chris Lattner
4be2e72cc5 revert recent patch which is causing widespread breakage. 
llvm-svn: 52415
 2008-06-17 17:06:43 +00:00
Evan Cheng
d43a533365 Don't forget to initialize SymbolSearchingDisabled. 
llvm-svn: 52414
 2008-06-17 16:49:02 +00:00
Duncan Sands
ec220a7c48 Fix typo that changed the logic to something wrong. 
Spotted by Nick Lewycky.

llvm-svn: 52411
 2008-06-17 15:55:30 +00:00
Duncan Sands
ae9d876e69 Split type expansion into ExpandInteger and ExpandFloat 
rather than bundling them together.  Rename FloatToInt
to PromoteFloat (better, if not perfect).  Reorganize
files by types rather than by operations.

llvm-svn: 52408
 2008-06-17 14:27:01 +00:00
Matthijs Kooijman
dd32cb58b5 Learn IPConstProp to propagate arguments that are directly returned. Strictly 
speaking these are not constant values. However, when a function always returns
one of its arguments, then from the point of view of each caller the return
value is constant (or at least a known value) and can be replaced.

llvm-svn: 52397
 2008-06-17 12:20:24 +00:00
Matthijs Kooijman
f0adaf34a1 Learn IPConstProp to look at individual return values and propagate them 
individually.

Also learn IPConstProp how returning first class aggregates work, in addition
to old style multiple return instructions.

Modify the return-constants testscase to confirm this behaviour.

llvm-svn: 52396
 2008-06-17 12:02:52 +00:00
Matthijs Kooijman
dd18157e57 Use a SmallVector instead of an array, since auto_ptr doesn't handle arrays 
properly.

llvm-svn: 52390
 2008-06-17 08:24:37 +00:00
Chris Lattner
f461cca6b7 add a new -enable-value-prop flag for llcbeta, that enables propagation 
of value info (sign/zero ext info) from one MBB to another.  This doesn't
handle much right now because of two limitations:

1) only handles zext/sext, not random bit propagation (no assert exists 
   for this)
2) doesn't handle phis.

llvm-svn: 52383
 2008-06-17 06:09:18 +00:00
Duncan Sands
915b66e08b Fix spelling. 
llvm-svn: 52381
 2008-06-17 03:24:13 +00:00
Evan Cheng
89e2e3292d Rather than avoiding to wrap ISD::DECLARE GV operand in X86ISD::Wrapper, simply handle it at dagisel time with x86 specific isel code. 
llvm-svn: 52377
 2008-06-17 02:01:22 +00:00
Chris Lattner
468adb09d7 Add a missing ~ (dtor became ctor) which caused crashes on a bunch of stuff. 
llvm-svn: 52374
 2008-06-16 23:06:51 +00:00
Evan Cheng
8cfd1d39a1 Do not issue identity copies. 
llvm-svn: 52373
 2008-06-16 22:52:53 +00:00
Dan Gohman
c1fd5f170b Refine the change in r52258 for avoiding use-before-def conditions 
when changing the stride of a comparison so that it's slightly
more precise, by having it scan the instruction list to determine
if there is a use of the condition after the point where the
condition will be inserted.

llvm-svn: 52371
 2008-06-16 22:34:15 +00:00
Chris Lattner
dd707dc82d switch TypeHasCycleThroughItself from using an std::set to using a SmallPtrSet, 
this speeds up the linking testcase in PR1860 by 44% (.379 -> 0.263)

llvm-svn: 52365
 2008-06-16 21:20:58 +00:00
Chris Lattner
6782f56054 stop making PATypeHolder's so crazily. 
llvm-svn: 52364
 2008-06-16 21:17:12 +00:00
Evan Cheng
4e7b7b21a2 Horizontal-add instructions are not commutative. 
llvm-svn: 52363
 2008-06-16 21:16:24 +00:00
Evan Cheng
7b87eb0128 Switch over to SetVector to ensure same order of iterations do not vary across runs. 
llvm-svn: 52361
 2008-06-16 21:08:17 +00:00
Chris Lattner
3dfc439a87 use a real associative container for type association instead of using 
a vector with a linear search.  This speeds up the linking testcase 
in PR1860 from 0.965s to 0.385s on my system.

llvm-svn: 52357
 2008-06-16 21:00:18 +00:00
Evan Cheng
acd614c262 mpsadbw is commutable. 
llvm-svn: 52352
 2008-06-16 20:25:59 +00:00
Chris Lattner
f87729aa1f bail out sooner if we have two concrete but different types. 
llvm-svn: 52351
 2008-06-16 20:03:01 +00:00
Chris Lattner
896c38c96a simplify some code. 
llvm-svn: 52350
 2008-06-16 19:55:40 +00:00
Chris Lattner
fe3ddf06a8 Apply a patch from Nathan Keynes, which speeds up llvm-link on 
the testcases in PR1860 from taking more than 1 hour (when I killed it)
to taking 1s.

llvm-svn: 52347
 2008-06-16 19:48:08 +00:00
Owen Anderson
63a604a316 Remove special case handling of empty MBBs now that we assign indices to them. 
llvm-svn: 52345
 2008-06-16 19:32:40 +00:00
Chris Lattner
1a604edf16 handle vectors. Any integers that got here would necessarily be different already. 
llvm-svn: 52341
 2008-06-16 18:27:53 +00:00
Chris Lattner
c24aba4724 Simplify ResolveTypes by pulling the null case out into the one 
client that cares and simplifying its control flow.

Remove the DestST argument to ResolveTypes and RecursiveResolveTypes*
which are dead now.

llvm-svn: 52340
 2008-06-16 18:19:05 +00:00
Evan Cheng
6c6518787b Iterating over SmallPtrSet is not deterministic. 
llvm-svn: 52339
 2008-06-16 18:17:09 +00:00
Chris Lattner
cebf7d79b3 simplify RecursiveResolveTypes and ResolveTypes by pulling the naming out of 
ResolveTypes into the one place that needs it.

llvm-svn: 52338
 2008-06-16 18:11:40 +00:00
Chris Lattner
b2c9317b8d Add a new flag that disables symbol lookup with dlsym when set. This allows 
a JIT client to completely control symbol lookup with the LazyFunctionCreator
interface.

llvm-svn: 52335
 2008-06-16 17:44:14 +00:00
Chris Lattner
ddfa0590ae Add support for icache invalidation on non-darwin ppc systems. 
Patch by Gary Benson!

llvm-svn: 52332
 2008-06-16 17:04:06 +00:00
Owen Anderson
3c4bb80d78 Re-enable empty block indexing by default, since it doesn't seem to have any 
impact on code quality or compile time.

llvm-svn: 52329
 2008-06-16 16:58:24 +00:00
Matthijs Kooijman
2ba9a750dd Make BuildSubAggregate use FindInsertedElement again to prevent it from 
inserting extractvalues. In particular, this prevents the insertion of
extractvalues that can't be folded away later. Also add an example of when this
stuff is needed.

llvm-svn: 52328
 2008-06-16 14:13:46 +00:00
Matthijs Kooijman
544c9b2fb0 Make the InsertBefore argument to FindInsertedValue optional, so you can find an inserted value without modifying the code. 
llvm-svn: 52319
 2008-06-16 13:28:31 +00:00
Matthijs Kooijman
238b1e8d69 Pass around Instruction* instead of Instruction& in FindInsertedValue and friends. 
llvm-svn: 52318
 2008-06-16 13:13:08 +00:00
Matthijs Kooijman
dedcf00fcc 80 column fixes. 
llvm-svn: 52316
 2008-06-16 12:57:37 +00:00
Matthijs Kooijman
1dd7d9cdc1 Move FindScalarValue from InstructionCombining.cpp to ValueTracking.cpp. While 
I'm at it, rename it to FindInsertedValue.

The only functional change is that newly created instructions are no longer
added to instcombine's worklist, but that is not really necessary anyway (and
I'll commit some improvements next that will completely remove the need).

llvm-svn: 52315
 2008-06-16 12:48:21 +00:00
Argyrios Kyrtzidis
65f671fd74 Minor comment fix. 
llvm-svn: 52312
 2008-06-16 10:14:09 +00:00
Duncan Sands
78bdcc813e Allow these transforms for types like i256 while 
still excluding types like i1 (not byte sized)
and i120 (loading an i120 requires loading an i64,
an i32, an i16 and an i8, which is expensive). 

llvm-svn: 52310
 2008-06-16 08:14:38 +00:00
Evan Cheng
93d91a771c Fix read after free found by valgrind. 
llvm-svn: 52309
 2008-06-16 07:34:17 +00:00
Evan Cheng
2dfe8c2435 Add option to commuteInstruction() which forces it to create a new (commuted) instruction. 
llvm-svn: 52308
 2008-06-16 07:33:11 +00:00
Owen Anderson
114ce854d1 Make indexing empty basic blocks an option for the moment. 
llvm-svn: 52306
 2008-06-16 07:10:49 +00:00
Chris Lattner
0bc9e4aa8f simplify some code by using a helper function. This really really 
wants a 'nocapture' predicate.

llvm-svn: 52304
 2008-06-16 06:38:26 +00:00
Chris Lattner
35be1f6c83 move a bunch of predicates up into their own section 
in this file, no other changes.

llvm-svn: 52303
 2008-06-16 06:30:22 +00:00
Chris Lattner
12453dd555 Other parts of this code treat noalias arguments as objects for 
the purposes of escape analysis.

llvm-svn: 52302
 2008-06-16 06:28:01 +00:00
Chris Lattner
e987a3bdd1 If we are checking to see if the result of a call aliases a 
pointer derived from a local allocation, if the local allocation
never escapes, the pointers can't alias.  This implements PR2436

llvm-svn: 52301
 2008-06-16 06:19:11 +00:00
Owen Anderson
1077c28bb1 Assign indices to empty basic blocks. This will be necessary for StrongPHIElimination in the near future. 
llvm-svn: 52300
 2008-06-16 06:18:41 +00:00
Chris Lattner
4e38446b34 Refactor basicaa's main alias function somethin' fierce. 
This fixes several minor bugs (such as returning noalias
for comparisons between external weak functions an null) but
is mostly a cleanup.

llvm-svn: 52299
 2008-06-16 06:10:11 +00:00
Chris Lattner
0b945db11c Switch from generating the int128 typedefs based on targetdata to generating 
them based on the end-compiler's capabilities.  This fixes PR2453

llvm-svn: 52297
 2008-06-16 04:25:29 +00:00
Chris Lattner
ac4fe3f51c Fix the crash on SimplifyLibCalls/2005-05-20-sprintf-crash.ll 
llvm-svn: 52295
 2008-06-16 04:10:21 +00:00
Chris Lattner
09554fd076 fix pr2460 
llvm-svn: 52294
 2008-06-16 04:02:40 +00:00
Duncan Sands
2dffe1cc15 The transforms in visitEXTRACT_VECTOR_ELT are 
not valid if the load is volatile.  Hopefully
all wrong DAG combiner transforms of volatile
loads and stores have now been caught.

llvm-svn: 52293
 2008-06-15 20:12:31 +00:00
Duncan Sands
2671518b8b LegalizeTypes support for INSERT_VECTOR_ELT with 
a non-constant index.

llvm-svn: 52292
 2008-06-15 20:00:14 +00:00
Wojciech Matyjewicz
58f5cc8864 Fix PR2434. When scanning for exising binary operator to reuse don't 
take into account the instrucion pointed by InsertPt. Thanks to it, 
returning the new value of InsertPt to the InsertBinop() caller can be 
avoided. The bug was, actually, in visitAddRecExpr() method which wasn't 
correctly handling changes of InsertPt. There shouldn't be any 
performance regression, as -gvn pass (run after -indvars) removes any 
redundant binops.

llvm-svn: 52291
 2008-06-15 19:07:39 +00:00
Wojciech Matyjewicz
0f03682725 Add a missing semicolon. 
llvm-svn: 52290
 2008-06-15 18:02:47 +00:00
Argyrios Kyrtzidis
6f81329b15 Fix the sys::Path::getSuffix() implementation. 
llvm-svn: 52288
 2008-06-15 15:15:19 +00:00
Argyrios Kyrtzidis
121cf54264 Fix the environment block that is passed to the CreateProcess function. 
This bug made llvm-ld unable to function with "-native" option, since the process that was used to call 'gcc' was crashing.

llvm-svn: 52284
 2008-06-15 03:54:39 +00:00
Duncan Sands
fa6e02c4dc Remove a redundant AfterLegalize check. Turn 
on some code when !AfterLegalize - but since
this whole code section is turned off by an
"if (0)" it's not really turning anything on.

llvm-svn: 52276
 2008-06-14 17:48:34 +00:00
Wojciech Matyjewicz
f66a7fe9de Change 'while' loop to 'do' loop. 
Add a safety measure. It isn't safe to assume in ScalarEvolutionExpander that
all loops are in canonical form (but it should be safe for loops that have
AddRecs).
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

llvm-svn: 52275
 2008-06-14 16:48:22 +00:00
Andrew Lenharth
327c3e7559 add missing atomic intrinsic from gcc 
llvm-svn: 52270
 2008-06-14 05:48:15 +00:00
Chris Lattner
f9d43c9853 Fix a case where tailcallelim wouldn't set the changed bit when it made a change. 
llvm-svn: 52267
 2008-06-14 00:49:48 +00:00
Evan Cheng
2e99c9cbf8 Teach the spiller to commute instructions in order to fold a reload. This hits 410 times on 444.namd and 122 times on 252.eon. 
llvm-svn: 52266
 2008-06-13 23:58:02 +00:00
Eli Friedman
11d4c94933 Don't skip over instructions other than loads that might read memory 
when trying to sink stores.

llvm-svn: 52259
 2008-06-13 22:02:12 +00:00
Dan Gohman
7f1fcc4339 Protect ChangeCompareStride from situations in which it is possible 
for it to generate use-before-def IR, such as in this testcase.

llvm-svn: 52258
 2008-06-13 21:43:41 +00:00
Eli Friedman
d38a639deb Make sure SimplifyStoreAtEndOfBlock doesn't mess with loops; the 
structure checks are incorrect if the blocks aren't distinct.
Fixes PR2435.

llvm-svn: 52257
 2008-06-13 21:17:49 +00:00
Duncan Sands
40c8db881a Disable some DAG combiner optimizations that may be 
wrong for volatile loads and stores.  In fact this
is almost all of them!  There are three types of
problems: (1) it is wrong to change the width of
a volatile memory access.  These may be used to
do memory mapped i/o, in which case a load can have
an effect even if the result is not used.  Consider
loading an i32 but only using the lower 8 bits.  It
is wrong to change this into a load of an i8, because
you are no longer tickling the other three bytes.  It
is also unwise to make a load/store wider.  For
example, changing an i16 load into an i32 load is
wrong no matter how aligned things are, since the
fact of loading an additional 2 bytes can have
i/o side-effects.  (2) it is wrong to change the
number of volatile load/stores: they may be counted
by the hardware.  (3) it is wrong to change a volatile
load/store that requires one memory access into one
that requires several.  For example on x86-32, you
can store a double in one processor operation, but to
store an i64 requires two (two i32 stores).  In a
multi-threaded program you may want to bitcast an i64
to a double and store as a double because that will
occur atomically, and be indivisible to other threads.
So it would be wrong to convert the store-of-double
into a store of an i64, because this will become two
i32 stores - no longer atomic.  My policy here is
to say that the number of processor operations for
an illegal operation is undefined.  So it is alright
to change a store of an i64 (requires at least two
stores; but could be validly lowered to memcpy for
example) into a store of double (one processor op).
In short, if the new store is legal and has the same
size then I say that the transform is ok.  It would
also be possible to say that transforms are always
ok if before they were illegal, whether after they
are illegal or not, but that's more awkward to do
and I doubt it buys us anything much.
However this exposed an interesting thing - on x86-32
a store of i64 is considered legal!  That is because
operations are marked legal by default, regardless of
whether the type is legal or not.  In some ways this
is clever: before type legalization this means that
operations on illegal types are considered legal;
after type legalization there are no illegal types
so now operations are only legal if they really are.
But I consider this to be too cunning for mere mortals.
Better to do things explicitly by testing AfterLegalize.
So I have changed things so that operations with illegal
types are considered illegal - indeed they can never
map to a machine operation.  However this means that
the DAG combiner is more conservative because before
it was "accidentally" performing transforms where the
type was illegal because the operation was nonetheless
marked legal.  So in a few such places I added a check
on AfterLegalize, which I suppose was actually just
forgotten before.  This causes the DAG combiner to do
slightly more than it used to, which resulted in the X86
backend blowing up because it got a slightly surprising
node it wasn't expecting, so I tweaked it.

llvm-svn: 52254
 2008-06-13 19:07:40 +00:00
Wojciech Matyjewicz
d3d353dd68 Use recently added getTruncateOrZeroExtend method to make the code shorter. 
llvm-svn: 52251
 2008-06-13 17:02:03 +00:00
Nick Lewycky
0aa53f2b52 Crash less. The i64 restriction in BinomialCoefficient caused some problems 
with code that was expecting different bit widths for different values.

Make getTruncateOrZeroExtend a method on ScalarEvolution, and use it.

llvm-svn: 52248
 2008-06-13 04:38:55 +00:00
Gabor Greif
10de8c6c59 fix a minor deviation from the original in my previous commit 
llvm-svn: 52247
 2008-06-12 21:51:29 +00:00
Gabor Greif
509b3a75f4 op_iterator-ify some loops, low hanging fruit only, there is more 
llvm-svn: 52246
 2008-06-12 21:37:33 +00:00
Evan Cheng
5897328890 Do not speculatively execute an instruction by hoisting it to its predecessor BB if any of its operands are defined but not used in BB. The transformation will prevent the operand from being sunk into the use block. 
llvm-svn: 52244
 2008-06-12 21:15:59 +00:00
Evan Cheng
b0d847cf05 Revert 52223. 
llvm-svn: 52243
 2008-06-12 20:55:39 +00:00
Owen Anderson
22a982f9eb Switch GVN to use ScopedHashTable. 
llvm-svn: 52242
 2008-06-12 19:25:32 +00:00
Matthijs Kooijman
43eae10af6 Fix redirection of stderr in sys::Program::ExecuteAndWait. There was logic 
error that caused it to redirect stderr to stdout too often.

This fix is applied identically to the win32 code as well, but that is
untested.
--Thi line, and those below, will be ignored--

M    System/Unix/Program.inc
M    System/Win32/Program.inc

llvm-svn: 52233
 2008-06-12 12:53:35 +00:00
Matthijs Kooijman
f2d854e34b Make I/O redirection handling in sys::Program a bit more consistent. No 
functional changes. Win32 code is untested, but should work fine.

In the unix variant, rename RedirectFD to RedirectIO and let that function
handle empty and null paths instead of doing that in the caller 3 times. This
is the same as win32 already does it.

In the win32 variant, use Path::isEmpty() instead of checking the resulting
c_str() manually. This is the same as unix already does it.

llvm-svn: 52230
 2008-06-12 10:47:18 +00:00
Gabor Greif
a90427821b op_iterator-ify some loops, fix 80col violations 
llvm-svn: 52226
 2008-06-11 21:38:51 +00:00
Anton Korobeynikov
74422b3cd0 Properly lower DYNAMIC_STACKALLOC - bracket all black magic with 
CALLSEQ_BEGIN & CALLSEQ_END.

llvm-svn: 52225
 2008-06-11 20:16:42 +00:00
Evan Cheng
7b7a321f64 For now, avoid generating FP select instructions in order to speculatively execute integer arithmetic instructions. FP selects are more likely to be expensive (even compared to branch on fcmp). This is not a wonderful solution but I rather err on the side of conservative. 
This fixes the heapsort performance regressions.

llvm-svn: 52224
 2008-06-11 19:18:20 +00:00
Evan Cheng
9b0c0a0f00 Avoid duplicating loop header which leads to unnatural loops (and just seem like general badness to me, likely to cause code explosion). 
Patch by Florian Brandner.

llvm-svn: 52223
 2008-06-11 19:07:54 +00:00
Matthijs Kooijman
0f9df32e12 Teach instruction combining about the extractvalue. It can succesfully fold 
useless insert-extract chains, similar to how it folds them for vectors.

Add a testcase for this.

llvm-svn: 52217
 2008-06-11 14:05:05 +00:00
Duncan Sands
2a9c84481c Sometimes (rarely) nodes held in LegalizeTypes 
maps can be deleted.  This happens when RAUW
replaces a node N with another equivalent node
E, deleting the first node.  Solve this by
adding (N, E) to ReplacedNodes, which is already
used to remap nodes to replacements.  This means
that deleted nodes are being allowed in maps,
which can be delicate: the memory may be reused
for a new node which might get confused with the
old deleted node pointer hanging around in the
maps, so detect this and flush out maps if it
occurs (ExpungeNode).  The expunging operation
is expensive, however it never occurs during
a llvm-gcc bootstrap or anywhere in the nightly
testsuite.  It occurs three times in "make check":
Alpha/illegal-element-type.ll,
PowerPC/illegal-element-type.ll and
X86/mmx-shift.ll.  If expunging proves to be too
expensive then there are other more complicated
ways of solving the problem.
In the normal case this patch adds the overhead
of a few more map lookups, which is hopefully
negligable.

llvm-svn: 52214
 2008-06-11 11:42:12 +00:00
Gordon Henriksen
22f7a6eb91 Better test for availability of __gnu_cxx::stdio_filebuf. 
If this doesn't work, I'll write a configure test.

llvm-svn: 52213
 2008-06-11 10:46:24 +00:00
Matthijs Kooijman
511d6a5cd3 Clarify a comment. 
llvm-svn: 52212
 2008-06-11 09:00:12 +00:00
Gabor Greif
519129edc8 op_iterator-ify loops 
llvm-svn: 52191
 2008-06-10 22:03:26 +00:00
Dan Gohman
c87fbbd014 Teach isGAPlusOffset to respect a GlobalAddressSDNode's offset 
value, which is something that apparently isn't used much.

llvm-svn: 52158
 2008-06-09 22:05:52 +00:00
Dan Gohman
68f8fbdac4 Re-apply 52002, allowing the verifier to accept non-MRV struct return 
types on functions, with adjustments so that it accepts both
new-style aggregate returns and old-style MRV returns, including those
with only a single member.

llvm-svn: 52157
 2008-06-09 21:26:13 +00:00
Dan Gohman
8c089d4df6 CodeGen support for aggregate-value function arguments. 
llvm-svn: 52156
 2008-06-09 21:19:23 +00:00
Duncan Sands
e46308480d Various tweaks related to apint codegen. No functionality 
change for non-funky-sized integers.

llvm-svn: 52151
 2008-06-09 15:48:25 +00:00
Dan Gohman
d789392934 Handle empty aggregate values. 
llvm-svn: 52150
 2008-06-09 15:21:47 +00:00
Dan Gohman
7c97681798 AsmParser support for immediate constant aggregate values. 
llvm-svn: 52149
 2008-06-09 14:45:02 +00:00
Dan Gohman
ce662293ed CPPBackend support for extractvalue and insertvalue. 
llvm-svn: 52147
 2008-06-09 14:12:10 +00:00
Dan Gohman
ee7b8797d0 Abort on an unrecognized opcode. 
llvm-svn: 52146
 2008-06-09 14:09:13 +00:00