RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-25 04:02:41 +01:00
Code Issues Projects Releases Wiki Activity
68,062 Commits 4 Branches 5 Tags 704 MiB
ea59381fc8
Commit Graph

7303 Commits

Author SHA1 Message Date
John Thompson
ae3a86d6de Added skeleton for inline asm multiple alternative constraint support. 
llvm-svn: 113766
 2010-09-13 18:15:37 +00:00
Owen Anderson
9c34a7831d Re-apply r113679, which was reverted in r113720, which added a paid of new instcombine transforms 
to expose greater opportunities for store narrowing in codegen.  This patch fixes a potential
infinite loop in instcombine caused by one of the introduced transforms being overly aggressive.

llvm-svn: 113763
 2010-09-13 17:59:27 +00:00
Eric Christopher
d4aaabfa74 Revert 113679, it was causing an infinite loop in a testcase that I've sent 
on to Owen.

llvm-svn: 113720
 2010-09-12 06:09:23 +00:00
Owen Anderson
d4ebde12ce Invert and-of-or into or-of-and when doing so would allow us to clear bits of the and's mask. 
This can result in increased opportunities for store narrowing in code generation.  Update a number of
tests for this change.  This fixes <rdar://problem/8285027>.

Additionally, because this inverts the order of ors and ands, some patterns for optimizing or-of-and-of-or
no longer fire in instances where they did originally.  Add a simple transform which recaptures most of these
opportunities: if we have an or-of-constant-or and have failed to fold away the inner or, commute the order 
of the two ors, to give the non-constant or a chance for simplification instead.

llvm-svn: 113679
 2010-09-11 05:48:06 +00:00
Gabor Greif
dfe6dea95f typoes 
llvm-svn: 113647
 2010-09-10 22:25:58 +00:00
Michael J. Spencer
98ad3f2ea7 CMake: Get rid of LLVMLibDeps.cmake and export the libraries normally. 
llvm-svn: 113632
 2010-09-10 21:14:25 +00:00
Benjamin Kramer
cd84f14639 This transform is also performed by InstructionSimplify, remove the duplicate. 
llvm-svn: 113608
 2010-09-10 19:52:35 +00:00
Owen Anderson
c0664eb99a Lower the unrolling theshold to 150. Empirical tests indicate that this is a sweet spot in the performance per 
code size increase curve.

llvm-svn: 113595
 2010-09-10 17:57:00 +00:00
Owen Anderson
f7276d42d3 What the loop unroller cares about, rather than just not unrolling loops with calls, is 
not unrolling loops that contain calls that would be better off getting inlined.  This mostly
comes up when an interleaved devirtualization pass has devirtualized a call which the inliner
will inline on a future pass.  Thus, rather than blocking all loops containing calls, add
a metric for "inline candidate calls" and block loops containing those instead.

llvm-svn: 113535
 2010-09-09 20:32:23 +00:00
Owen Anderson
db6a08beef Revert r113439, which relaxed the requirement that loops containing calls cannot be unrolled. After some discussion, 
there seems to be a better way to achieve the same effect.

llvm-svn: 113528
 2010-09-09 20:02:23 +00:00
Owen Anderson
82f5f82ccd r113526 introduced an unintended change to the loop unrolling threshold. Revert it. 
llvm-svn: 113527
 2010-09-09 19:11:57 +00:00
Owen Anderson
46189436ea Fix typo in code to cap the loop code size reduction calculation. 
llvm-svn: 113526
 2010-09-09 19:08:59 +00:00
Owen Anderson
a182e2b20b Use code-size reduction metrics to estimate the amount of savings we'll get when we unroll a loop. 
Next step is to recalculate the threshold values given this new heuristic.

llvm-svn: 113525
 2010-09-09 19:07:31 +00:00
Owen Anderson
956afdd1f2 Relax the "don't unroll loops containing calls" rule. Instead, when a loop contains a call, lower the 
unrolling threshold to the optimize-for-size threshold.  Basically, for loops containing calls, unrolling
can still be profitable as long as the loop is REALLY small.

llvm-svn: 113439
 2010-09-08 23:10:07 +00:00
Owen Anderson
c51d7d1a8d Generalize instcombine's support for combining multiple bit checks into a single test. Patch by Dirk Steinke! 
llvm-svn: 113423
 2010-09-08 22:16:17 +00:00
Owen Anderson
38493be0ce Add a separate unrolling threshold when the current function is being optimized for size. 
The threshold value of 50 is arbitrary, and I chose it simply by analogy to the inlining thresholds, where
the baseline unrolling threshold is slightly smaller than the baseline inlining threshold.  This could
undoubtedly use some tuning.

llvm-svn: 113306
 2010-09-07 23:15:30 +00:00
Chris Lattner
a58a97dafc Fix a serious performance regression introduced by r108687 on linux: 
turning (fptrunc (sqrt (fpext x))) -> (sqrtf x)  is great, but we have
to delete the original sqrt as well.  Not doing so causes us to do 
two sqrt's when building with -fmath-errno (the default on linux).

llvm-svn: 113260
 2010-09-07 20:01:38 +00:00
Nick Lewycky
8c155b346e Fix major bug in thunk detection. Also verify the calling convention. 
Switch from isWeakForLinker to mayBeOverridden which is more accurate.

Add more statistics and debugging info. Add comments. Move static function
outside anonymous namespace.

llvm-svn: 113190
 2010-09-07 01:42:10 +00:00
Chris Lattner
6e6a535055 fix PR8067, an over-aggressive assertion in LICM. 
llvm-svn: 113146
 2010-09-06 05:11:24 +00:00
Chris Lattner
4100881939 Teach loop rotate to hoist trivially invariant instructions 
in the duplicated block instead of duplicating them.  

Duplicating them into the end of the loop and the preheader 
means that we got a phi node in the header of the loop, 
which prevented LICM from hoisting them.  GVN would
usually come around later and merge the duplicated 
instructions so we'd get reasonable output... except that
anything dependent on the shoulda-been-hoisted value can't
be hoisted.  In PR5319 (which this fixes), a memory value
didn't get promoted.

llvm-svn: 113134
 2010-09-06 01:10:22 +00:00
Chris Lattner
818045f208 pull a simple method out of LICM into a new 
Loop::hasLoopInvariantOperands method. Remove
a useless and confusing Loop::isLoopInvariant(Instruction)
method, which didn't do what you thought it did.

No functionality change.

llvm-svn: 113133
 2010-09-06 01:05:37 +00:00
Chris Lattner
e54b9e2332 more cleanups 
llvm-svn: 113115
 2010-09-05 20:13:07 +00:00
Chris Lattner
41d0913a1d Change lower atomic pass to use IntrinsicInst to simplify it a bit. 
llvm-svn: 113114
 2010-09-05 20:10:47 +00:00
Chris Lattner
fc6bf7cb34 eliminate some non-obvious casts. UndefValue isa Constant. 
llvm-svn: 113113
 2010-09-05 20:03:09 +00:00
Nick Lewycky
d48f0c5718 Fix warning reported by MSVC++ builder. 
llvm-svn: 113106
 2010-09-05 09:11:38 +00:00
Nick Lewycky
97d03432c5 Switch FnSet to containing the ComparableFunction instead of a pointer to one. 
This reduces malloc traffic (yay!) and removes MergeFunctionsEqualityInfo.

llvm-svn: 113105
 2010-09-05 09:00:32 +00:00
Nick Lewycky
b505af7ec4 Fix many bugs when merging weak-strong and weak-weak pairs. We now merge all 
strong functions first to make sure they're the canonical definitions and then
do a second pass looking only for weak functions.

llvm-svn: 113104
 2010-09-05 08:22:49 +00:00
Chris Lattner
8df3ffd7ac zap dead code. 
llvm-svn: 113073
 2010-09-04 18:12:00 +00:00
Dan Gohman
e1ad0ebbcc Fix LoopSimplify to notify ScalarEvolution when splitting a loop backedge 
into an inner loop, as the new loop iteration may differ substantially.
This fixes PR8078.

llvm-svn: 113057
 2010-09-04 02:42:48 +00:00
Chris Lattner
2b77a2a167 fix a bug in my licm rewrite when a load from the promoted memory 
location is being re-stored to the memory location.  We would get
a dangling pointer from the SSAUpdate data structure and miss a 
use.  This fixes PR8068

llvm-svn: 113042
 2010-09-04 00:12:30 +00:00
Owen Anderson
94d98b12c8 Propagate non-local comparisons. Fixes PR1757. 
llvm-svn: 113025
 2010-09-03 22:47:08 +00:00
Owen Anderson
9161c79ffe Add support for simplifying a load from a computed value to a load from a global when it 
is provable that they're equivalent.  This fixes PR4855.

llvm-svn: 112994
 2010-09-03 19:08:37 +00:00
Chris Lattner
1edeb00c72 fix more AST updating bugs, correcting miscompilation in PR8041 
llvm-svn: 112878
 2010-09-02 22:19:10 +00:00
Duncan Sands
a95548a8d2 Reapply commit 112699, speculatively reverted by echristo, since 
I'm sure it is harmless.  Original commit message:
If PrototypeValue is erased in the middle of using the SSAUpdator
then the SSAUpdator may access freed memory.  Instead, simply pass
in the type and name explicitly, which is all that was used anyway.

llvm-svn: 112810
 2010-09-02 08:14:03 +00:00
Chris Lattner
b1c861e28c deepen my MMX/SRoA hack to avoid hurting non-x86 codegen. 
llvm-svn: 112763
 2010-09-01 23:09:27 +00:00
Dan Gohman
d0dc80485c Fix loop unswitching's assumption that a code path which either 
infinite loops or exits will eventually exit. This fixes PR5373.

llvm-svn: 112745
 2010-09-01 21:46:45 +00:00
Owen Anderson
e036dd5b0b JumpThreading keeps LazyValueInfo up to date, so we don't need to rerun it 
if we schedule another LVI-using pass afterwards.

llvm-svn: 112722
 2010-09-01 18:27:22 +00:00
Eric Christopher
f72b7040a2 Speculatively revert 112699 and 112702, they seem to be causing 
self host errors on clang-x86-64.

llvm-svn: 112719
 2010-09-01 17:29:10 +00:00
Duncan Sands
f7e2f45e1a If PrototypeValue is erased in the middle of using the SSAUpdator 
then the SSAUpdator may access freed memory.  Instead, simply pass
in the type and name explicitly, which is all that was used anyway.

llvm-svn: 112699
 2010-09-01 10:29:33 +00:00
Chris Lattner
9759e898f0 add a gross hack to work around a problem that Argiris reported 
on llvmdev: SRoA is introducing MMX datatypes like <1 x i64>,
which then cause random problems because the X86 backend is
producing mmx stuff without inserting proper emms calls.

In the short term, force off MMX datatypes.  In the long term,
the X86 backend should not select generic vector types to MMX
registers.  This is being worked on, but won't be done in time
for 2.8.  rdar://8380055

llvm-svn: 112696
 2010-09-01 05:14:33 +00:00
Dan Gohman
ad593f9d93 Revert 112442 and 112440 until the compile time problems introduced 
by 112440 are resolved.

llvm-svn: 112692
 2010-09-01 01:45:53 +00:00
Chris Lattner
5ed5e29575 licm is wasting time hoisting constant foldable operations, 
instead of hoisting them, just fold them away.  This occurs in the
testcase for PR8041, for example.

llvm-svn: 112669
 2010-08-31 23:00:16 +00:00
Chris Lattner
f2ee739a2b tidy up 
llvm-svn: 112643
 2010-08-31 21:21:25 +00:00
Owen Anderson
233463074b More cleanups of my JumpThreading transforms, including extracting some duplicated code into a helper function. 
llvm-svn: 112634
 2010-08-31 20:26:04 +00:00
Owen Anderson
66b51ff843 Add an RAII helper to make cleanup of the RecursionSet more fool-proof. 
llvm-svn: 112628
 2010-08-31 19:24:27 +00:00
Owen Anderson
5e2c04e417 Only try to clean up the current block if we changed that block already. 
llvm-svn: 112625
 2010-08-31 18:55:52 +00:00
Owen Anderson
e2b5bd3a7f Refactor my fix for PR5652 to terminate the predecessor lookups after the first failure. 
llvm-svn: 112620
 2010-08-31 18:48:48 +00:00
Nick Lewycky
66fe124a92 Fix an infinite loop; merging two functions will create a new function (if the 
two are weak, we make them thunks to a new strong function) so don't iterate
through the function list as we're modifying it.

Also add back the outermost loop which got removed during the cleanups.

llvm-svn: 112595
 2010-08-31 08:29:37 +00:00
Owen Anderson
bf12defee5 Don't perform an extra traversal of the function just to do cleanup. We can safely simplify instructions after each block has been processed without worrying about iterator invalidation. 
llvm-svn: 112594
 2010-08-31 07:55:56 +00:00
Owen Anderson
ccaee65189 Rename ValuePropagation to a more descriptive CorrelatedValuePropagation. 
llvm-svn: 112591
 2010-08-31 07:48:34 +00:00
Owen Anderson
6853ce863c Rename file to something more descriptive. 
llvm-svn: 112590
 2010-08-31 07:41:39 +00:00
Owen Anderson
ba28fe3dcb More Chris-inspired JumpThreading fixes: use ConstantExpr to correctly constant-fold undef, and be more careful with its return value. 
This actually exposed an infinite recursion bug in ComputeValueKnownInPredecessors which theoretically already existed (in JumpThreading's
handling of and/or of i1's), but never manifested before.  This patch adds a tracking set to prevent this case.

llvm-svn: 112589
 2010-08-31 07:36:34 +00:00
Nick Lewycky
75dfadbaf9 Switch to DenseSet, simplifying much more code. We now have a single iteration 
where we hash, compare and fold, instead of one iteration where we build up
the hash buckets and a second one to fold.

llvm-svn: 112582
 2010-08-31 05:53:05 +00:00
Owen Anderson
bd9edea8a3 Remove r111665, which implemented store-narrowing in InstCombine. Chris discovered a miscompilation in it, and it's not easily 
fixable at the optimizer level. I'll investigate reimplementing it in DAGCombine.

llvm-svn: 112575
 2010-08-31 04:41:06 +00:00
Owen Anderson
f1cec75012 Fix a typo. 
llvm-svn: 112560
 2010-08-30 23:59:30 +00:00
Owen Anderson
c90a98e0a5 Cleanups suggested by Chris. 
llvm-svn: 112553
 2010-08-30 23:34:17 +00:00
Owen Anderson
af33f22b40 Re-apply r112539, being more careful to respect the return values of the constant folding methods. Additionally, 
use the ConstantExpr::get*() methods to simplify some constant folding.

llvm-svn: 112550
 2010-08-30 23:22:36 +00:00
Owen Anderson
8e72b2f3f5 Add statistics to evaluate this pass. 
llvm-svn: 112545
 2010-08-30 22:45:55 +00:00
Owen Anderson
9d301e20e8 Revert r112539. It accidentally introduced a miscompilation. 
llvm-svn: 112543
 2010-08-30 22:33:41 +00:00
Owen Anderson
479c0c406f Fixes and cleanups pointed out by Chris. In general, be careful to handle 0 results from ComputeValueKnownInPredecessors 
(indicating undef), and re-use existing constant folding APIs.

llvm-svn: 112539
 2010-08-30 22:07:52 +00:00
Chris Lattner
92879a5ba1 rewrite DwarfEHPrepare to use SSAUpdater to promote its allocas 
instead of PromoteMemToReg.  This allows it to stop using DF and DT,
eliminating a computation of DT and DF from clang -O3.  Clang is now
down to 2 runs of DomFrontier.

llvm-svn: 112457
 2010-08-29 19:54:28 +00:00
Chris Lattner
9e35d96cea two changes: 1) make AliasSet hold the list of call sites with an 
assertingvh so we get a violent explosion if the pointer dangles.

2) Fix AliasSetTracker::deleteValue to remove call sites with
   by-pointer comparisons instead of by-alias queries.  Using
   findAliasSetForCallSite can cause alias sets to get merged
   when they shouldn't, and can also miss alias sets when the
   call is readonly.

#2 fixes PR6889, which only repros with a .c file :(

llvm-svn: 112452
 2010-08-29 18:42:23 +00:00
Chris Lattner
51639dea34 LICM does get dead instructions input to it. Instead of sinking them 
out of loops, just delete them.

llvm-svn: 112451
 2010-08-29 18:22:25 +00:00
Chris Lattner
aac9263929 use moveBefore instead of remove+insert, it avoids some 
symtab manipulation, so its faster (in addition to being
more elegant)

llvm-svn: 112450
 2010-08-29 18:18:40 +00:00
Chris Lattner
5c31ee4663 revert 112448 for now. 
llvm-svn: 112449
 2010-08-29 18:11:16 +00:00
Chris Lattner
8e922b170c optimize LICM::hoist to use moveBefore. Correct its updating 
of AST to remove the hoisted instruction from the AST, since it
is no longer in the loop.

llvm-svn: 112448
 2010-08-29 18:03:33 +00:00
Chris Lattner
150bdce5c1 fix some bugs (found by inspection) where LICM would not update 
LICM correctly.  When sinking an instruction, it should not add
entries for the sunk instruction to the AST, it should remove
the entry for the sunk instruction.  The blocks being sunk to
are not in the loop, so their instructions shouldn't be in the
AST (yet)!

llvm-svn: 112447
 2010-08-29 18:00:00 +00:00
Chris Lattner
a3b10c0752 rework the ownership of subloop alias information: instead of 
keeping them around until the pass is destroyed, keep them
around a) just when useful (not for outer loops) and b) destroy
them right after we use them.  This should reduce memory use
and fixes potential bugs where a loop is deleted and another
loop gets allocated to the same address.

llvm-svn: 112446
 2010-08-29 17:46:00 +00:00
Chris Lattner
4cdad46980 apparently unswitch had the same "Feature". Stop its 
claims that it preserves domfrontier if it doesn't really.

llvm-svn: 112445
 2010-08-29 17:23:19 +00:00
Chris Lattner
7181337888 now that loop passes don't use DomFrontier, there is no reason 
for the unroller to pretend it supports updating it.  It still
has a horrible hack for DomTree.

llvm-svn: 112444
 2010-08-29 17:21:35 +00:00
Dan Gohman
d7ad97b6f5 Optionally rerun dedicated-register filtering after applying 
other filtering techniques, as those may allow it to filter
out more obviously unprofitable candidates.

llvm-svn: 112441
 2010-08-29 16:39:22 +00:00
Dan Gohman
d5e0d45b20 Fix several areas in LSR to do a better job keeping the main 
LSRInstance data structures up to date. This fixes some
pessimizations caused by stale data which will be exposed
in an upcoming change.

llvm-svn: 112440
 2010-08-29 16:32:54 +00:00
Dan Gohman
b62347f067 Refactor the three main groups of code out of 
NarrowSearchSpaceUsingHeuristics into separate functions.

llvm-svn: 112439
 2010-08-29 16:09:42 +00:00
Dan Gohman
7b45482eab Delete a bogus check. 
llvm-svn: 112438
 2010-08-29 15:30:29 +00:00
Dan Gohman
bde89495fa Add some comments. 
llvm-svn: 112437
 2010-08-29 15:27:08 +00:00
Dan Gohman
185c024c53 Move this debug output into GenerateAllReuseFormula, to declutter 
the high-level logic.

llvm-svn: 112436
 2010-08-29 15:21:38 +00:00
Dan Gohman
5023e9f408 Delete an unused declaration. 
llvm-svn: 112435
 2010-08-29 15:19:11 +00:00
Dan Gohman
defdc9d59c Do one lookup instead of two. 
llvm-svn: 112434
 2010-08-29 15:18:49 +00:00
Chris Lattner
270d50d48c licm preserves the cfg, it doesn't have to explicitly say it 
preserves domfrontier.  It does preserve AA though.

llvm-svn: 112419
 2010-08-29 07:02:56 +00:00
Chris Lattner
40ec41d1c4 now that it doesn't use the PromoteMemToReg function, LICM doesn't 
require DomFrontier.  Dropping this doesn't actually save any runs
of the pass though.

llvm-svn: 112418
 2010-08-29 06:49:44 +00:00
Chris Lattner
fe3e4cdd30 completely rewrite the memory promotion algorithm in LICM. 
Among other things, this uses SSAUpdater instead of 
PromoteMemToReg.

llvm-svn: 112417
 2010-08-29 06:43:52 +00:00
Chris Lattner
9921d9c3c1 use getUniqueExitBlocks instead of a manual set. 
llvm-svn: 112412
 2010-08-29 05:12:21 +00:00
Chris Lattner
2133f877c6 reimplement LICM::sink to use SSAUpdater instead of PromoteMemToReg. 
This leads to much simpler code.

llvm-svn: 112410
 2010-08-29 04:55:06 +00:00
Chris Lattner
fac07b1dca implement SSAUpdater::RewriteUseAfterInsertions, a helpful form of RewriteUse. 
llvm-svn: 112409
 2010-08-29 04:54:06 +00:00
Chris Lattner
24927beaff remove dead proto 
llvm-svn: 112408
 2010-08-29 04:53:24 +00:00
Chris Lattner
65ce6da2f1 reduce indentation in LICM::sink by using early exits, use 
getUniqueExitBlocks instead of getExitBlocks and a manual
set to eliminate dupes.

llvm-svn: 112405
 2010-08-29 04:28:20 +00:00
Chris Lattner
4928fe010e modernize this pass a bit: use efficient set/map and reduce indentation. 
llvm-svn: 112404
 2010-08-29 04:23:04 +00:00
Chris Lattner
ecf276b787 remove unions from LLVM IR. They are severely buggy and not 
being actively maintained, improved, or extended.

llvm-svn: 112356
 2010-08-28 04:09:24 +00:00
Chris Lattner
4b49ada02c remove the ABCD and SSI passes. They don't have any clients that 
I'm aware of, aren't maintained, and LVI will be replacing their value.
nlewycky approved this on irc.

llvm-svn: 112355
 2010-08-28 03:51:24 +00:00
Chris Lattner
fc1da78d16 for completeness, allow undef also. 
llvm-svn: 112351
 2010-08-28 03:36:51 +00:00
Chris Lattner
b2dbdbc795 squish dead code. 
llvm-svn: 112350
 2010-08-28 03:21:03 +00:00
Chris Lattner
b61cf1e296 handle the constant case of vector insertion. For something 
like this:

struct S { float A, B, C, D; };

struct S g;
struct S bar() { 
  struct S A = g;
  ++A.B;
  A.A = 42;
  return A;
}

we now generate:

_bar:                                   ## @bar
## BB#0:                                ## %entry
	movq	_g@GOTPCREL(%rip), %rax
	movss	12(%rax), %xmm0
	pshufd	$16, %xmm0, %xmm0
	movss	4(%rax), %xmm2
	movss	8(%rax), %xmm1
	pshufd	$16, %xmm1, %xmm1
	unpcklps	%xmm0, %xmm1
	addss	LCPI1_0(%rip), %xmm2
	pshufd	$16, %xmm2, %xmm2
	movss	LCPI1_1(%rip), %xmm0
	pshufd	$16, %xmm0, %xmm0
	unpcklps	%xmm2, %xmm0
	ret

instead of:

_bar:                                   ## @bar
## BB#0:                                ## %entry
	movq	_g@GOTPCREL(%rip), %rax
	movss	12(%rax), %xmm0
	pshufd	$16, %xmm0, %xmm0
	movss	4(%rax), %xmm2
	movss	8(%rax), %xmm1
	pshufd	$16, %xmm1, %xmm1
	unpcklps	%xmm0, %xmm1
	addss	LCPI1_0(%rip), %xmm2
	movd	%xmm2, %eax
	shlq	$32, %rax
	addq	$1109917696, %rax       ## imm = 0x42280000
	movd	%rax, %xmm0
	ret

llvm-svn: 112345
 2010-08-28 01:50:57 +00:00
Chris Lattner
c70b0c0ee7 optimize bitcasts from large integers to vector into vector 
element insertion from the pieces that feed into the vector.
This handles a pattern that occurs frequently due to code
generated for the x86-64 abi.  We now compile something like
this:

struct S { float A, B, C, D; };
struct S g;
struct S bar() { 
  struct S A = g;
  ++A.A;
  ++A.C;
  return A;
}

into all nice vector operations:

_bar:                                   ## @bar
## BB#0:                                ## %entry
	movq	_g@GOTPCREL(%rip), %rax
	movss	LCPI1_0(%rip), %xmm1
	movss	(%rax), %xmm0
	addss	%xmm1, %xmm0
	pshufd	$16, %xmm0, %xmm0
	movss	4(%rax), %xmm2
	movss	12(%rax), %xmm3
	pshufd	$16, %xmm2, %xmm2
	unpcklps	%xmm2, %xmm0
	addss	8(%rax), %xmm1
	pshufd	$16, %xmm1, %xmm1
	pshufd	$16, %xmm3, %xmm2
	unpcklps	%xmm2, %xmm1
	ret

instead of icky integer operations:

_bar:                                   ## @bar
	movq	_g@GOTPCREL(%rip), %rax
	movss	LCPI1_0(%rip), %xmm1
	movss	(%rax), %xmm0
	addss	%xmm1, %xmm0
	movd	%xmm0, %ecx
	movl	4(%rax), %edx
	movl	12(%rax), %esi
	shlq	$32, %rdx
	addq	%rcx, %rdx
	movd	%rdx, %xmm0
	addss	8(%rax), %xmm1
	movd	%xmm1, %eax
	shlq	$32, %rsi
	addq	%rax, %rsi
	movd	%rsi, %xmm1
	ret

This resolves rdar://8360454

llvm-svn: 112343
 2010-08-28 01:20:38 +00:00
Benjamin Kramer
edb09ef2df Update CMake build. Add newline at end of file. 
llvm-svn: 112332
 2010-08-28 00:11:12 +00:00
Owen Anderson
dc4703bcd5 Add a prototype of a new peephole optimizing pass that uses LazyValue info to simplify PHIs and select's. 
This pass addresses the missed optimizations from PR2581 and PR4420.

llvm-svn: 112325
 2010-08-27 23:31:36 +00:00
Chris Lattner
3f880c2097 Enhance the shift propagator to handle the case when you have: 
A = shl x, 42
...
B = lshr ..., 38

which can be transformed into:
A = shl x, 4
...

iff we can prove that the would-be-shifted-in bits
are already zero.  This eliminates two shifts in the testcase
and allows eliminate of the whole i128 chain in the real example.

llvm-svn: 112314
 2010-08-27 22:53:44 +00:00
Chris Lattner
80632e5fd9 Implement a pretty general logical shift propagation 
framework, which is good at ripping through bitfield
operations.  This generalize a bunch of the existing
xforms that instcombine does, such as 
  (x << c) >> c -> and
to handle intermediate logical nodes.  This is useful for
ripping up the "promote to large integer" code produced by
SRoA.

llvm-svn: 112304
 2010-08-27 22:24:38 +00:00
Chris Lattner
5ed3d56ced remove some special shift cases that have been subsumed into the 
more general simplify demanded bits logic.

llvm-svn: 112291
 2010-08-27 21:04:34 +00:00
Owen Anderson
a1a80a3acd Fix typos in comments. 
llvm-svn: 112286
 2010-08-27 20:32:56 +00:00
Chris Lattner
866b888095 teach the truncation optimization that an entire chain of 
computation can be truncated if it is fed by a sext/zext that doesn't
have to be exactly equal to the truncation result type.

llvm-svn: 112285
 2010-08-27 20:32:06 +00:00
Chris Lattner
69a9143584 Add an instcombine to clean up a common pattern produced 
by the SRoA "promote to large integer" code, eliminating
some type conversions like this:

   %94 = zext i16 %93 to i32                       ; <i32> [#uses=2]
   %96 = lshr i32 %94, 8                           ; <i32> [#uses=1]
   %101 = trunc i32 %96 to i8                      ; <i8> [#uses=1]

This also unblocks other xforms from happening, now clang is able to compile:

struct S { float A, B, C, D; };
float foo(struct S A) { return A.A + A.B+A.C+A.D; }

into:

_foo:                                   ## @foo
## BB#0:                                ## %entry
	pshufd	$1, %xmm0, %xmm2
	addss	%xmm0, %xmm2
	movdqa	%xmm1, %xmm3
	addss	%xmm2, %xmm3
	pshufd	$1, %xmm1, %xmm0
	addss	%xmm3, %xmm0
	ret

on x86-64, instead of:

_foo:                                   ## @foo
## BB#0:                                ## %entry
	movd	%xmm0, %rax
	shrq	$32, %rax
	movd	%eax, %xmm2
	addss	%xmm0, %xmm2
	movapd	%xmm1, %xmm3
	addss	%xmm2, %xmm3
	movd	%xmm1, %rax
	shrq	$32, %rax
	movd	%eax, %xmm0
	addss	%xmm3, %xmm0
	ret

This seems pretty close to optimal to me, at least without
using horizontal adds.  This also triggers in lots of other
code, including SPEC.

llvm-svn: 112278
 2010-08-27 18:31:05 +00:00
Owen Anderson
35ff7a208e Use LVI to eliminate conditional branches where we've tested a related condition previously. Update tests for this change. 
This fixes PR5652.

llvm-svn: 112270
 2010-08-27 17:12:29 +00:00
Chris Lattner
d5d68438c1 optimize "integer extraction out of the middle of a vector" as produced 
by SRoA.  This is part of rdar://7892780, but needs another xform to
expose this.

llvm-svn: 112232
 2010-08-26 22:14:59 +00:00
Chris Lattner
19a5dc488b optimize bitcast(trunc(bitcast(x))) where the result is a float and 'x' 
is a vector to be a vector element extraction.  This allows clang to
compile:

struct S { float A, B, C, D; };
float foo(struct S A) { return A.A + A.B+A.C+A.D; }

into:

_foo:                                   ## @foo
## BB#0:                                ## %entry
	movd	%xmm0, %rax
	shrq	$32, %rax
	movd	%eax, %xmm2
	addss	%xmm0, %xmm2
	movapd	%xmm1, %xmm3
	addss	%xmm2, %xmm3
	movd	%xmm1, %rax
	shrq	$32, %rax
	movd	%eax, %xmm0
	addss	%xmm3, %xmm0
	ret

instead of:

_foo:                                   ## @foo
## BB#0:                                ## %entry
	movd	%xmm0, %rax
	movd	%eax, %xmm0
	shrq	$32, %rax
	movd	%eax, %xmm2
	addss	%xmm0, %xmm2
	movd	%xmm1, %rax
	movd	%eax, %xmm1
	addss	%xmm2, %xmm1
	shrq	$32, %rax
	movd	%eax, %xmm0
	addss	%xmm1, %xmm0
	ret

... eliminating half of the horribleness.

llvm-svn: 112227
 2010-08-26 21:55:42 +00:00
Owen Anderson
77fcf53657 Make JumpThreading smart enough to properly thread StrSwitch when it's compiled with clang++. 
llvm-svn: 112198
 2010-08-26 17:40:24 +00:00
Dan Gohman
8088d5e31d Reapply r112091 and r111922, support for metadata linking, with a 
fix: add a flag to MapValue and friends which indicates whether
any module-level mappings are being made. In the common case of
inlining, no module-level mappings are needed, so MapValue doesn't
need to examine non-function-local metadata, which can be very
expensive in the case of a large module with really deep metadata
(e.g. a large C++ program compiled with -g).

This flag is a little awkward; perhaps eventually it can be moved
into the ClonedCodeInfo class.

llvm-svn: 112190
 2010-08-26 15:41:53 +00:00
Daniel Dunbar
3839de70f5 Revert r111922, "MapValue support for MDNodes. This is similar to r109117, 
except ...", it is causing *massive* performance regressions when building Clang
with itself (-O3 -g).

llvm-svn: 112158
 2010-08-26 03:48:11 +00:00
Daniel Dunbar
aeb8abb0e0 Revert r112091, "Remap metadata attached to instructions when remapping 
individual ...", which depends on r111922, which I am reverting.

llvm-svn: 112157
 2010-08-26 03:48:08 +00:00
Chris Lattner
164c35930a zap dead code. 
llvm-svn: 112130
 2010-08-26 01:13:54 +00:00
Dan Gohman
c7605a66b7 Rewrite ExtractGV, removing a bunch of stuff that didn't fully work, 
and was over-complicated, and replacing it with a simple implementation.

llvm-svn: 112120
 2010-08-26 00:22:55 +00:00
Chris Lattner
ef3055ca05 remove some llvmcontext arguments that are now dead post-refactoring. 
llvm-svn: 112104
 2010-08-25 23:00:45 +00:00
Dan Gohman
d19a0a49d1 Remap metadata attached to instructions when remapping individual 
instructions, not when remapping modules.

llvm-svn: 112091
 2010-08-25 21:36:50 +00:00
Devang Patel
05becf3ac5 DIGlobalVariable can be used to encode debug info for globals that are directly folded into a constant by FE. 
llvm-svn: 112072
 2010-08-25 18:52:02 +00:00
Dan Gohman
30fd721e48 Use MapValue in the Linker instead of having a private function 
which does the same thing. This eliminates redundant code and
handles MDNodes better. MDNode linking still doesn't fully
work yet though.

llvm-svn: 111941
 2010-08-24 18:50:07 +00:00
Owen Anderson
a4d73e6be9 Turn LVI on, previously detected failures should be fixed now. 
llvm-svn: 111923
 2010-08-24 17:21:18 +00:00
Dan Gohman
5cbd5888e7 MapValue support for MDNodes. This is similar to r109117, except 
that it avoids a lot of unnecessary cloning by avoiding remapping
MDNode cycles when none of the nodes in the cycle actually need to
be remapped. Also it uses the new temporary MDNode mechanism.

llvm-svn: 111922
 2010-08-24 17:10:10 +00:00
Owen Anderson
9f7621fa94 Turn LVI back off, I have a testcase now. 
llvm-svn: 111834
 2010-08-23 19:59:27 +00:00
Owen Anderson
ee69c0112d Re-enable LazyValueInfo. Monitoring for failures. 
llvm-svn: 111816
 2010-08-23 18:12:23 +00:00
Owen Anderson
fe3d206e65 Now that PassInfo and Pass::ID have been separated, move the rest of the passes over to the new registration API. 
llvm-svn: 111815
 2010-08-23 17:52:01 +00:00
Owen Anderson
678fd04aa5 Re-apply r111568 with a fix for the clang self-host. 
llvm-svn: 111665
 2010-08-20 18:24:43 +00:00
Owen Anderson
0e57acb623 Revert r111568 to unbreak clang self-host. 
llvm-svn: 111571
 2010-08-19 23:25:16 +00:00
Owen Anderson
7f2852ba2d When a set of bitmask operations, typically from a bitfield initialization, only modifies the low bytes of a value, 
we can narrow the store to only over-write the affected bytes.

llvm-svn: 111568
 2010-08-19 22:15:40 +00:00
Owen Anderson
e22f313e44 Disable LVI while I evaluate a failure. 
llvm-svn: 111551
 2010-08-19 19:47:08 +00:00
Owen Anderson
9de85fb6c0 Tentatively enabled LVI by default. I'll be monitoring for any failures. 
llvm-svn: 111543
 2010-08-19 19:04:40 +00:00
Dan Gohman
ba5736d2fe Process the step before the start, because it's usually the simpler 
of the two.

llvm-svn: 111495
 2010-08-19 01:02:31 +00:00
Owen Anderson
65795241db Inform LazyValueInfo whenever a block is deleted, to avoid dangling pointer issues. 
llvm-svn: 111382
 2010-08-18 18:39:01 +00:00
Chris Lattner
ab876b6ce8 Fix PR7755: knowing something about an inval for a pred 
from the LHS should disable reconsidering that pred on the
RHS.  However, knowing something about the pred on the RHS
shouldn't disable subsequent additions on the RHS from
happening.

llvm-svn: 111349
 2010-08-18 03:14:36 +00:00
Chris Lattner
4c6e9192d1 fit in 80 cols 
llvm-svn: 111348
 2010-08-18 03:13:35 +00:00
Chris Lattner
d20a060467 remove some dead code. 
llvm-svn: 111344
 2010-08-18 02:41:56 +00:00
Chris Lattner
686cddf177 remove dead prototype. 
llvm-svn: 111342
 2010-08-18 02:37:06 +00:00
Eric Christopher
08e9f0250a Temporarily revert r110987 as it's causing some miscompares in 
vector heavy code.  I'll re-enable when we've tracked down the problem.

llvm-svn: 111318
 2010-08-17 22:55:27 +00:00
Dan Gohman
e26025ddd0 When rotating loops, put the original header at the bottom of the 
loop, making the resulting loop significantly less ugly.  Also, zap
its trivial PHI nodes, since it's easy.

llvm-svn: 111255
 2010-08-17 17:39:21 +00:00
Dan Gohman
aec27f6889 Use the getUniquePredecessor() utility function, instead of doing 
what it does manually.

llvm-svn: 111248
 2010-08-17 17:07:02 +00:00
Evan Cheng
908b65c371 Add an option to disable codegen prepare critical edge splitting. In theory, PHI elimination is already doing all (most?) of the splitting needed. But machine-licm and machine-sink seem to miss some important optimizations when splitting is disabled. 
llvm-svn: 111224
 2010-08-17 01:34:49 +00:00
Dan Gohman
7900e1ace3 Instead of having CollectSubexpr's categorize operands as interesting or 
uninteresting, just put all the operands on one list and make
GenerateReassociations make the decision about what's interesting.
This is simpler, and it avoids an extra ScalarEvolution::getAddExpr call.

llvm-svn: 111133
 2010-08-16 15:50:00 +00:00
Dan Gohman
38d11cdfe0 Put add operands in ScalarEvolution-canonical order, when convenient. 
This isn't necessary, because ScalarEvolution sorts them anyway,
but it's tidier this way.

llvm-svn: 111132
 2010-08-16 15:39:27 +00:00
Dan Gohman
80b2503100 Avoid #include <ScalarEvolution.h> in LoopSimplify.cpp, which doesn't 
actually use ScalarEvolution.

llvm-svn: 111124
 2010-08-16 14:44:03 +00:00
Dan Gohman
9178d0792f Instead, teach SimplifyCFG to trim non-address-taken blocks from 
indirectbr destination lists.

llvm-svn: 111122
 2010-08-16 14:41:14 +00:00
Dan Gohman
afb3db46d2 LoopSimplify shouldn't split loop backedges that use indirectbr. PR7867. 
llvm-svn: 111061
 2010-08-14 00:43:09 +00:00
Dan Gohman
d04a608a73 Teach SimplifyCFG how to simplify indirectbr instructions. 
- Eliminate redundant successors.
 - Convert an indirectbr with one successor into a direct branch.

Also, generalize SimplifyCFG to be able to be run on a function entry block.
It knows quite a few simplifications which are applicable to the entry
block, and it only needs a few checks to avoid trouble with the entry block.

llvm-svn: 111060
 2010-08-14 00:29:42 +00:00
Dan Gohman
3e7c2a2040 Fix LSR's ExtractImmediate and ExtractSymbol to avoid calling 
ScalarEvolution::getAddExpr, which can be pretty expensive, when nothing
has changed, which is pretty common.

llvm-svn: 111042
 2010-08-13 21:17:19 +00:00
Nate Begeman
e57074fc48 Reapply this transformation now that it is passing the external test which it previously failed. 
llvm-svn: 110987
 2010-08-13 00:17:53 +00:00
Chris Lattner
fd40059e71 fix PR7876: If ipsccp decides that a function's address is taken 
before it rewrites the code, we need to use that in the post-rewrite pass.

llvm-svn: 110962
 2010-08-12 22:25:23 +00:00
Eric Christopher
34acdf57df Temporarily revert 110737 and 110734, they were causing failures 
in an external testsuite.

llvm-svn: 110905
 2010-08-12 07:01:22 +00:00
Nate Begeman
713062e756 Add the minimal amount of smarts necessary to instcombine of shufflevectors to recognize 
patterns generated by clang for transpose of a matrix in generic vectors.  This is made
of two parts:

1) Propagating vector extracts of hi/lo half into their users
2) Recognizing an insertion of even elements followed by the odd elements as an unpack.

Testcase to come, but this shrinks the # of shuffle instructions generated on x86 from ~40 to the minimal 8.

llvm-svn: 110734
 2010-08-10 21:38:12 +00:00
Nick Lewycky
920eab5cc1 Fix a use after free error caught by the valgrind builders. 
llvm-svn: 110601
 2010-08-09 21:03:28 +00:00
Eli Friedman
7197d66ff1 PR7853: fix a silly mistake introduced in r101899, and add a test to make sure 
it doesn't regress again.

llvm-svn: 110597
 2010-08-09 20:49:43 +00:00
Nick Lewycky
50705ca94a Do more to modernize MergeFunctions. Refactor in response to Chris' code review. 
llvm-svn: 110538
 2010-08-08 05:04:23 +00:00
Owen Anderson
fb2c32c524 Don't attempt the PRE inline asm calls, since we don't value number them yet. Fixes PR7835. 
llvm-svn: 110489
 2010-08-07 00:20:35 +00:00
Dan Gohman
13c66ce161 Eliminate PromoteMemoryToRegisterID; just use addPreserved("mem2reg") 
instead, as an example of what this looks like.

llvm-svn: 110478
 2010-08-06 21:48:06 +00:00
Owen Anderson
f2fea95f2f Reapply r110396, with fixes to appease the Linux buildbot gods. 
llvm-svn: 110460
 2010-08-06 18:33:48 +00:00
Nick Lewycky
8c4f9777de Fix uninitialized variable warning. 
Also move 'default' case next to a real case to help compiler optimize in
non-Debug builds.
No functionality change.

llvm-svn: 110435
 2010-08-06 07:43:46 +00:00
Nick Lewycky
178f0d4cd0 Work in progress, cleaning up MergeFuncs. 
Further clean up the comparison function by removing overly generalized
"domains".
Remove all understanding of ELF aliases and simplify folding code and comments.

llvm-svn: 110434
 2010-08-06 07:21:30 +00:00
Owen Anderson
aadd8a89ca Revert r110396 to fix buildbots. 
llvm-svn: 110410
 2010-08-06 00:23:35 +00:00
Owen Anderson
b9762c07cb Don't use PassInfo* as a type identifier for passes. Instead, use the address of the static 
ID member as the sole unique type identifier.  Clean up APIs related to this change.

llvm-svn: 110396
 2010-08-05 23:42:04 +00:00
Owen Anderson
bbeb4e2e58 Give JumpThreading+LVI a long-form cl::opt so that it's easier to toggle the default. 
llvm-svn: 110384
 2010-08-05 22:11:31 +00:00
Owen Anderson
0c1fe93d7c Experiments show that we can safely increase our unrolling threshold without unduly impacting code size, particularly 
since unrolling is not enabled at -Os.

llvm-svn: 110233
 2010-08-04 18:32:46 +00:00
Dan Gohman
88d90617de Fix whitespace. 
llvm-svn: 110223
 2010-08-04 17:43:57 +00:00
Dan Gohman
bea5c99638 Fix a comment. 
llvm-svn: 110181
 2010-08-04 01:16:35 +00:00
Dan Gohman
a087b90950 Thread const correctness through a bunch of AliasAnalysis interfaces and 
eliminate several const_casts.

Make CallSite implicitly convertible to ImmutableCallSite.

Rename the getModRefBehavior for intrinsic IDs to
getIntrinsicModRefBehavior to avoid overload ambiguity with CallSite,
which happens to be implicitly convertible to bool.

llvm-svn: 110155
 2010-08-03 21:48:53 +00:00
Dan Gohman
a80f89dbc7 Make instcombine set explicit alignments on load or store 
instructions with alignment 0, so that subsequent passes don't
need to bother checking the TargetData ABI size manually.

llvm-svn: 110128
 2010-08-03 18:20:32 +00:00
Peter Collingbourne
10c4f9d6bd Add an atomic lowering pass 
llvm-svn: 110113
 2010-08-03 16:19:16 +00:00
Dan Gohman
8391bdbe95 Use unary + instead of a separate local variable for working 
around std::min vs static const friction.

llvm-svn: 110112
 2010-08-03 16:15:50 +00:00
Owen Anderson
e957c57ebb Re-apply the infamous r108614, with a fix pointed out by Dirk Steinke. 
llvm-svn: 110036
 2010-08-02 09:32:13 +00:00
Oscar Fuentes
4742c01c2a Prefix next' iterator operation with llvm::'. 
Fixes potential ambiguity problems on VS 2010.

Patch by nobled!

llvm-svn: 110029
 2010-08-02 06:00:15 +00:00
Daniel Dunbar
06a1d9ddcf Fix a -Wreorder warning. 
llvm-svn: 110022
 2010-08-02 05:43:46 +00:00
Nick Lewycky
747b2ba2b9 Work in progress. 
Start cleaning up MergeFunctions to look more like the rest of LLVM. The
primary change here is to move the methods responsible for comparison into the
new FunctionComparator object. Some comments added. There's more to do.

llvm-svn: 110021
 2010-08-02 05:23:03 +00:00
Daniel Dunbar
f2be238c99 Speculatively revert r108614, "Another attempt at getting the clang self-host to 
like my instcombine patch.", in an attempt to fix Clang i386 bootstrap.
 - Also PR7719.

llvm-svn: 109953
 2010-07-31 19:51:11 +00:00
Rafael Espindola
75734bc143 The BlockExtractorPass() constructor was not reading the BlockFile and that was 
exactly what bugpoint expected it to do.

There was also only one user of
BlockExtractorPass(const std::vector<BasicBlock*> &B), so just remove it and
make BlockExtractorPass read BlockFile.

This fixes bugpoint's block extraction.

Nick, please review.

llvm-svn: 109936
 2010-07-31 00:32:17 +00:00
Dan Gohman
230d92b375 Move MaximumAlignment to be a member of the Value class. 
llvm-svn: 109891
 2010-07-30 21:07:05 +00:00
Nick Lewycky
b6db4aaeaa Add missing newline to debug statement. 
llvm-svn: 109886
 2010-07-30 20:27:01 +00:00
Eli Friedman
642f4c66c3 PR7750: !CExpr->isNullValue() only properly computes whether CExpr is nonnull 
if CExpr is a ConstantInt.

llvm-svn: 109773
 2010-07-29 18:03:33 +00:00
Gabor Greif
0899f01287 simplify by using CallSite constructors; virtually eliminates CallSite::get from the tree 
llvm-svn: 109687
 2010-07-28 22:50:26 +00:00
Dan Gohman
939744be5f Define a maximum supported alignment value for load, store, and 
alloca instructions (constrained by their internal encoding),
and add error checking for it. Fix an instcombine bug which
generated huge alignment values (null is infinitely aligned).
This fixes undefined behavior noticed by John Regehr.

llvm-svn: 109643
 2010-07-28 20:12:04 +00:00
Dan Gohman
5295c21189 When user code intentionally dereferences null, the alignment of the 
dereference is theoretically infinite. Put a cap on the computed
alignment to avoid overflow, noticed by John Regehr.

llvm-svn: 109596
 2010-07-28 17:14:23 +00:00
Gabor Greif
ecd9b45b05 simplify 
llvm-svn: 109589
 2010-07-28 15:52:43 +00:00
Gabor Greif
16a98fddef use Value* constructor of CallSite to create potentially improper site, and test that 
llvm-svn: 109581
 2010-07-28 14:28:18 +00:00
Gabor Greif
2778a21978 recommit simplification (r109502, backed out r109509); seems to innocent 
llvm-svn: 109510
 2010-07-27 16:44:23 +00:00
Gabor Greif
c95e84ab61 back out this too to restore the bots 
llvm-svn: 109509
 2010-07-27 15:56:07 +00:00
Gabor Greif
394841cea8 simplify: CallSite::get --> CallSite constructor 
llvm-svn: 109506
 2010-07-27 15:02:37 +00:00
Gabor Greif
ee90e05139 simplify 
llvm-svn: 109502
 2010-07-27 13:31:22 +00:00
Owen Anderson
12ab6162bf Add an initial implementation of LazyValueInfo updating for JumpThreading. Disabled for now. 
llvm-svn: 109424
 2010-07-26 18:48:03 +00:00
Dan Gohman
23a6a1bdc5 Remove LCSSA's bogus dependence on LoopSimplify and LoopSimplify's bogus 
dependence on DominanceFrontier. Instead, add an explicit DominanceFrontier
pass in StandardPasses.h to ensure that it gets scheduled at the right
time.

Declare that loop unrolling preserves ScalarEvolution, and shuffle some
getAnalysisUsages.

This eliminates one LoopSimplify and one LCCSA run in the standard
compile opts sequence.

llvm-svn: 109413
 2010-07-26 18:11:16 +00:00
Dan Gohman
7d588b7dec Preserve ScalarEvolution in the loop unroller. 
llvm-svn: 109412
 2010-07-26 18:02:06 +00:00
Dan Gohman
32010a6c96 Use DominatorTree::properlyDominates instead of dominates with an 
explicit inequality check.

llvm-svn: 109401
 2010-07-26 17:37:36 +00:00
Dan Gohman
bed2730c5b A block dominates itself, by definition. 
llvm-svn: 109400
 2010-07-26 17:35:32 +00:00
Nick Lewycky
8cbc57da86 Revert this because we can't clone cyclic MDNodes which are creating during a 
build of llvm-gcc.

llvm-svn: 109355
 2010-07-24 20:54:02 +00:00
Nick Lewycky
7094bd27bc Whether function-local or not, a MDNode may reference a Function in which case 
it needs to be mapped to refer to the function in the new module, not the old
one. Fixes PR7700.

llvm-svn: 109353
 2010-07-24 19:43:25 +00:00
Devang Patel
f53de95e64 Speculatively revert 109117 
llvm-svn: 109132
 2010-07-22 18:44:00 +00:00
Gabor Greif
feb96f2df2 keep in 80 cols 
llvm-svn: 109122
 2010-07-22 17:18:03 +00:00
Devang Patel
58b7143697 Map MDNode correctly. 
A non function local MDNode can have an operand which is cloned by MapValue(). 

llvm-svn: 109117
 2010-07-22 16:35:00 +00:00
Gabor Greif
96a9f8c7c6 mass elimination of reliance on automatic iterator dereferencing 
llvm-svn: 109103
 2010-07-22 13:36:47 +00:00
Gabor Greif
225d763b90 simplify 
llvm-svn: 109101
 2010-07-22 13:07:39 +00:00
Gabor Greif
40e372be7e do not access arguments via low-level interface, do not multiply dereference use_iterators 
llvm-svn: 109100
 2010-07-22 13:04:32 +00:00
Gabor Greif
582eb39c11 pass dereferenced iterator to dyn_cast 
llvm-svn: 109099
 2010-07-22 11:48:35 +00:00
Gabor Greif
87ef691878 pass dereferenced iterator to dyn_cast 
llvm-svn: 109098
 2010-07-22 11:43:44 +00:00
Gabor Greif
a7509fca78 undo 80 column trespassing I caused 
llvm-svn: 109092
 2010-07-22 10:37:47 +00:00
Dan Gohman
859ffd353e Make NamedMDNode not be a subclass of Value, and simplify the interface 
for creating and populating NamedMDNodes.

llvm-svn: 109061
 2010-07-21 23:38:33 +00:00
Owen Anderson
f8addbb0a1 Fix batch of converting RegisterPass<> to INTIALIZE_PASS(). 
llvm-svn: 109045
 2010-07-21 22:09:45 +00:00
Dan Gohman
a07df2e2bf Make this code a little more readable. 
llvm-svn: 108968
 2010-07-20 23:49:44 +00:00
Dan Gohman
6f009c4a42 Use DebugLocs instead of MDNodes. 
llvm-svn: 108967
 2010-07-20 23:49:05 +00:00
Dan Gohman
c813346166 Fix a typo. 
llvm-svn: 108962
 2010-07-20 23:10:36 +00:00
Dan Gohman
66ae369e76 Don't look up the "dbg" metadata kind by name. 
llvm-svn: 108961
 2010-07-20 23:09:34 +00:00
Dan Gohman
d8e1fd258f Use getDebugLoc and setDebugLoc instead of getDbgMetadata and setDbgMetadata, 
avoiding MDNode overhead.

llvm-svn: 108909
 2010-07-20 20:09:07 +00:00
Dan Gohman
e36ed050b8 Remember that the induction variable is always a PHINode and 
use getIncomingValueForBlock instead of
LoopInfo::getCanonicalInductionVariableIncrement.

llvm-svn: 108865
 2010-07-20 17:18:52 +00:00
Owen Anderson
0217b8e1d9 Tweak per Chris' comments. 
llvm-svn: 108736
 2010-07-19 19:23:32 +00:00
Owen Anderson
8159755faa Reimplement r108639 in InstCombine rather than DAGCombine. 
llvm-svn: 108687
 2010-07-19 08:09:34 +00:00
Owen Anderson
c8dc055b5e Another attempt at getting the clang self-host to like my instcombine patch. 
llvm-svn: 108614
 2010-07-17 06:56:35 +00:00
Chris Lattner
663dd0eb36 eliminate unlockedRefineAbstractTypeTo, types are all per-llvmcontext, 
so there is no locking involved in type refinement.

llvm-svn: 108553
 2010-07-16 20:50:13 +00:00
Dan Gohman
b5593f8fda Reorder the contents of various getAnalysisUsage functions, eliminating 
a redundant loopsimplify run from the default -O2 sequence.

llvm-svn: 108539
 2010-07-16 17:58:45 +00:00
Owen Anderson
ca2a306a1a Remove the rest of my instcombine changes. Back to the drawing board on this one. 
llvm-svn: 108530
 2010-07-16 16:39:00 +00:00
Gabor Greif
17c48ecd68 eliminate CallInst::ArgOffset 
llvm-svn: 108522
 2010-07-16 09:38:02 +00:00
Nick Lewycky
1b4a83430b Arrays and vectors with different numbers of elements are not equivalent. 
llvm-svn: 108517
 2010-07-16 06:31:12 +00:00
Eric Christopher
5eef314caf Also revert 108422, it's causing some test failures. 
Working on testcases for Owen.

llvm-svn: 108494
 2010-07-16 01:36:12 +00:00
Dan Gohman
de662ad8b0 Don't merge uses when they are targetting fixup sites with 
different widths. In a use with a narrower fixup, formulae
may be wider than the fixup, in which case the high bits
aren't necessarily meaningful, so it isn't safe to reuse
them for uses with wider fixups.

This fixes PR7618, though the testcase is too large for a
reasonable regression test, since it heavily dependes on
hitting LSR's heuristics in a certain way.

llvm-svn: 108455
 2010-07-15 20:24:58 +00:00
Dan Gohman
3c5c24ac18 Use dbgs() instead of errs() in a DEBUG. 
llvm-svn: 108453
 2010-07-15 20:12:42 +00:00
Owen Anderson
cc5ef967da Speculatively revert r108429 to fix the clang self-host. 
llvm-svn: 108436
 2010-07-15 18:18:57 +00:00
Owen Anderson
565ac2758c Per Chris' suggestion, get rid of the select canonicalization and just add 
the corresponding or-icmp-and pattern.  This has the added benefit of doing
the matching earlier, and thus being less susceptible to being confused by
earlier transforms.

llvm-svn: 108429
 2010-07-15 17:24:23 +00:00
Owen Anderson
b9f325bab4 Remove unneeded check, and correct style. 
llvm-svn: 108427
 2010-07-15 16:38:22 +00:00
Dan Gohman
d75ba463e0 Watch out for a constant offset cancelling out a base register, forming 
a zero. This situation arrises in Fortran code with induction variables
that start at 1 instead of 0. This fixes PR7651.

llvm-svn: 108424
 2010-07-15 15:14:45 +00:00
Owen Anderson
01a2992a91 Reapply r108378, with bugfixes, testcase, and improved comment formatting. 
This now passes LIT, nighty test, and llvm-gcc bootstrap on my machine.

llvm-svn: 108422
 2010-07-15 15:00:23 +00:00
Nick Lewycky
933c0402bb This is a full sentence. 
llvm-svn: 108418
 2010-07-15 06:51:22 +00:00
Nick Lewycky
c4e3206edc Disable aliases on all platforms. 
llvm-svn: 108417
 2010-07-15 06:48:56 +00:00
Chris Lattner
63cfa4b8e0 make various clients of ReplaceAndSimplifyAllUses tolerate 
it *changing* the things it replaces, not just causing them
to drop to null.  There is no functionality change yet, but 
this is required for a subsequent patch.

llvm-svn: 108414
 2010-07-15 06:06:04 +00:00
Eli Friedman
ded29cd2dc Speculatively revert r108378; may be causing bootstrap failures. 
llvm-svn: 108389
 2010-07-15 00:33:00 +00:00
Owen Anderson
4656b749f2 Add instcombine transforms to optimize tests of multiple bits of the same value into a single larger comparison. 
llvm-svn: 108378
 2010-07-14 23:33:51 +00:00
Owen Anderson
5de6c42ebd Extend SimplifyCFG's common-destination folding heuristic to allow a single 
"bonus" instruction to be speculatively executed.  Add a heuristic to
ensure we're not tripping up out-of-order execution by checking that this bonus
instruction only uses values that were already guaranteed to be available.

This allows us to eliminate the short circuit in (x&1)&&(x&2).

llvm-svn: 108351
 2010-07-14 19:52:16 +00:00
Chris Lattner
38e6ecd9f1 revert r108320, I see the failures now... 
llvm-svn: 108322
 2010-07-14 06:16:35 +00:00
Chris Lattner
5822d6d579 reapply benjamin's instcombine patch, I don't see anything wrong with it and can't repro any problems with a manual self-host. 
llvm-svn: 108320
 2010-07-14 05:59:13 +00:00
Eric Christopher
f57bd9c94d Grammar. 
llvm-svn: 108252
 2010-07-13 18:27:13 +00:00
Duncan Sands
8864383748 Handle the case of a tail recursion in which the tail call is followed 
by a return that returns a constant, while elsewhere in the function
another return instruction returns a different constant.  This is a
special case of accumulator recursion, so just generalize the existing
logic a bit.

llvm-svn: 108241
 2010-07-13 15:41:41 +00:00
Benjamin Kramer
cf8ad46899 Nope, still breaks the release selfhost bots :( 
llvm-svn: 108153
 2010-07-12 16:38:48 +00:00
Benjamin Kramer
e391789246 Reapply the "or" half of r108136, which seems to be less problematic. 
llvm-svn: 108152
 2010-07-12 16:15:48 +00:00
Gabor Greif
9f6b0059e4 cache result of operator* 
llvm-svn: 108150
 2010-07-12 15:48:26 +00:00
Benjamin Kramer
98c95e7743 Revert r108141 again, sigh. 
llvm-svn: 108148
 2010-07-12 14:42:04 +00:00
Gabor Greif
04d1ca9e95 cache result of operator* 
llvm-svn: 108147
 2010-07-12 14:15:58 +00:00
Gabor Greif
b2b6867582 cache result of operator* 
llvm-svn: 108146
 2010-07-12 14:15:10 +00:00
Gabor Greif
bf2470e6e9 cache result of operator* 
llvm-svn: 108145
 2010-07-12 14:14:03 +00:00
Gabor Greif
de1c22125f cache result of operator* 
llvm-svn: 108144
 2010-07-12 14:13:15 +00:00
Gabor Greif
cf2cac8ef4 cache results of operator* 
llvm-svn: 108143
 2010-07-12 14:12:11 +00:00
Gabor Greif
8159329bfb cache results of operator* 
llvm-svn: 108142
 2010-07-12 14:10:24 +00:00
Benjamin Kramer
c4f46375d3 Reapply 108136 with an ugly pasto fixed. 
llvm-svn: 108141
 2010-07-12 13:44:00 +00:00
Benjamin Kramer
e12b88fafe Move optimization to avoid redundant matching. 
llvm-svn: 108140
 2010-07-12 13:34:22 +00:00
Benjamin Kramer
d9bf737e62 Revert r108136 until I figure out why it broke selfhost. 
llvm-svn: 108139
 2010-07-12 12:35:49 +00:00
Gabor Greif
8f93852b68 cache dereferenced iterators 
llvm-svn: 108138
 2010-07-12 12:03:02 +00:00
Gabor Greif
fc666b2d76 recommit r108131 (hich has been backed out in r108135) with a fix 
llvm-svn: 108137
 2010-07-12 12:02:10 +00:00
Benjamin Kramer
f00a49ceff instcombine: fold (x & y) | (~x & z) and (x & y) ^ (~x & z) into ((y ^ z) & x) ^ z which is one instruction shorter. (PR6773) 
before:
  %and = and i32 %y, %x
  %neg = xor i32 %x, -1
  %and4 = and i32 %z, %neg
  %xor = xor i32 %and4, %and

after:
  %xor1 = xor i32 %z, %y
  %and2 = and i32 %xor1, %x
  %xor = xor i32 %and2, %z

llvm-svn: 108136
 2010-07-12 11:54:45 +00:00
Gabor Greif
1ae3047b14 back out r108131 (of TailDuplication.cpp) for now, it causes a buildbot failure 
llvm-svn: 108135
 2010-07-12 11:32:39 +00:00
Gabor Greif
f420d43bf6 cache dereferenced iterators 
llvm-svn: 108134
 2010-07-12 11:19:24 +00:00
Gabor Greif
95805d51b8 cache dereferenced iterators 
llvm-svn: 108133
 2010-07-12 10:59:23 +00:00