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
57,230 Commits 4 Branches 5 Tags 704 MiB
c954fc3333
Commit Graph

1289 Commits

Author SHA1 Message Date
Chris Lattner
9c797fdf1d Speed up codegen prepare from 3.58s to 0.488s. 
llvm-svn: 96081
 2010-02-13 05:01:14 +00:00
Chris Lattner
975da20868 PHINode::getBasicBlockIndex is O(n) in the number of inputs 
to a PHI, avoid it in the common case where the BB occurs
in the same index for multiple phis.  This speeds up CGP on
an insane testcase from 8.35 to 3.58s.

llvm-svn: 96080
 2010-02-13 04:24:19 +00:00
Chris Lattner
1a067d53b0 iterate over preds using PHI information when available instead of 
using pred_begin/end.  It is much faster.

llvm-svn: 96079
 2010-02-13 04:15:26 +00:00
Jakob Stoklund Olesen
670458b3be Teach SimplifyCFG about magic pointer constants. 
Weird code sometimes uses pointer constants other than null. This patch
teaches SimplifyCFG to build switch instructions in those cases.

Code like this:

void f(const char *x) {
  if (!x)
    puts("null");
  else if ((uintptr_t)x == 1)
    puts("one");
  else if (x == (char*)2 || x == (char*)3)
    puts("two");
  else if ((intptr_t)x == 4)
    puts("four");
  else
    puts(x);
}

Now becomes a switch:

define void @f(i8* %x) nounwind ssp {
entry:
  %magicptr23 = ptrtoint i8* %x to i64            ; <i64> [#uses=1]
  switch i64 %magicptr23, label %if.else16 [
    i64 0, label %if.then
    i64 1, label %if.then2
    i64 2, label %if.then9
    i64 3, label %if.then9
    i64 4, label %if.then14
  ]

Note that LLVM's own DenseMap uses magic pointers.

llvm-svn: 95439
 2010-02-05 22:03:18 +00:00
Dan Gohman
8142ba6bbb Use a SmallSetVector instead of a SetVector; this code showed up as a 
malloc caller in a profile.

llvm-svn: 95407
 2010-02-05 19:20:15 +00:00
Bob Wilson
8207d33d94 Fix pr6198 by moving the isSized() check to an outer conditional. 
The testcase from pr6198 does not crash for me -- I don't know what's up with
that -- so I'm not adding it to the tests.

llvm-svn: 94984
 2010-02-01 17:41:44 +00:00
Bob Wilson
0f04082970 Check alignment of loads when deciding whether it is safe to execute them 
unconditionally.  Besides checking the offset, also check that the underlying
object is aligned as much as the load itself.

llvm-svn: 94875
 2010-01-30 04:42:39 +00:00
Bob Wilson
bb651db10d Use uint64_t instead of unsigned for offsets and sizes. 
llvm-svn: 94835
 2010-01-29 20:34:28 +00:00
Bob Wilson
f897b7b37e Improve isSafeToLoadUnconditionally to recognize that GEPs with constant 
indices are safe if the result is known to be within the bounds of the
underlying object.

llvm-svn: 94829
 2010-01-29 19:19:08 +00:00
Victor Hernandez
60060d030c mem2reg erases the dbg.declare intrinsics that it converts to dbg.val intrinsics 
llvm-svn: 94763
 2010-01-29 00:01:35 +00:00
Bob Wilson
2e1a609654 Avoid creating redundant PHIs in SSAUpdater::GetValueInMiddleOfBlock. 
This was already being done in SSAUpdater::GetValueAtEndOfBlock so I've
just changed SSAUpdater to check for existing PHIs in both places.

llvm-svn: 94690
 2010-01-27 22:01:02 +00:00
Benjamin Kramer
cfe07fadc7 Don't bother with sprintf, just pass the Twine through. 
llvm-svn: 94684
 2010-01-27 19:58:47 +00:00
Chris Lattner
deabd1cfb6 some cleanups. 
llvm-svn: 94649
 2010-01-27 02:12:20 +00:00
Chris Lattner
354782f9b9 no need to check for null 
llvm-svn: 94648
 2010-01-27 02:04:20 +00:00
Victor Hernandez
e6321dc910 When converting dbg.declare to dbg.value, attach promoted store's debug metadata to dbg.value 
llvm-svn: 94634
 2010-01-27 00:44:36 +00:00
Victor Hernandez
4e7937f77b Avoid extra calls to MD->getNumOperands() 
llvm-svn: 94618
 2010-01-26 23:29:09 +00:00
Victor Hernandez
3a9845ecf7 Switch AllocaDbgDeclares to SmallVector and don't leak DIFactory 
llvm-svn: 94567
 2010-01-26 18:57:53 +00:00
Victor Hernandez
f12e8a120f In mem2reg, for all alloca/stores that get promoted where the alloca has an associated llvm.dbg.declare instrinsic, insert an llvm.dbg.var intrinsic before each store. 
llvm-svn: 94493
 2010-01-26 02:42:15 +00:00
Victor Hernandez
253c09eb8f Revert r94260 until findDbgDeclare() is made more efficient 
llvm-svn: 94432
 2010-01-25 17:52:13 +00:00
Chris Lattner
5a57121631 make -fno-rtti the default unless a directory builds with REQUIRES_RTTI. 
llvm-svn: 94378
 2010-01-24 20:43:08 +00:00
Victor Hernandez
7ce9006ba0 In mem2reg, for all alloca/stores that get promoted where the alloca has an associated llvm.dbg.declare instrinsic, insert an llvm.dbg.var intrinsic before each store 
llvm-svn: 94260
 2010-01-23 00:17:34 +00:00
Victor Hernandez
e4d8ef8fc4 Keep ignoring pointer-to-pointer bitcasts 
llvm-svn: 94194
 2010-01-22 19:05:05 +00:00
Chris Lattner
276811b58a Stop building RTTI information for *most* llvm libraries. Notable 
missing ones are libsupport, libsystem and libvmcore.  libvmcore is
currently blocked on bugpoint, which uses EH.  Once it stops using
EH, we can switch it off.

This #if 0's out 3 unit tests, because gtest requires RTTI information.
Suggestions welcome on how to fix this.

llvm-svn: 94164
 2010-01-22 06:49:46 +00:00
Victor Hernandez
728ab49966 No need to look through bitcasts for DbgInfoIntrinsic 
llvm-svn: 94114
 2010-01-21 23:09:12 +00:00
Victor Hernandez
ef9e66c909 DbgInfoIntrinsic no longer appear in an instruction's use list 
llvm-svn: 94113
 2010-01-21 23:08:36 +00:00
Victor Hernandez
d44671b931 DbgInfoIntrinsics no longer appear in an instruction's use list; so clean up looking for them in use iterations and remove OnlyUsedByDbgInfoIntrinsics() 
llvm-svn: 94111
 2010-01-21 23:05:53 +00:00
Chris Lattner
642049e127 It turns out that this #include is needed because otherwise 
ValueMapper.cpp ends up calling an out of line 
__ZNK4llvm12PATypeHolder3getEv, which is a template and llvm-config
determines arbitrarily to use the one in libipo.  This sucks, but
keeping the #include is a reasonable workaround.

llvm-svn: 94103
 2010-01-21 21:29:25 +00:00
Chris Lattner
3547371b0b unbreak the build, apparently without this transformutils starts depending on libipa? 
llvm-svn: 94102
 2010-01-21 21:20:51 +00:00
Chris Lattner
3f91babac8 tidy up 
llvm-svn: 94101
 2010-01-21 21:05:54 +00:00
Victor Hernandez
507d86fcbb Don't need to include IntrinsicInst.h any more 
llvm-svn: 94092
 2010-01-21 19:33:59 +00:00
Victor Hernandez
3a31114cf9 No need to map NULL operands of metadata 
llvm-svn: 94091
 2010-01-21 19:26:20 +00:00
Victor Hernandez
50e7dcdcc2 Switch Elts from vector to SmallVector 
llvm-svn: 93989
 2010-01-20 06:56:16 +00:00
Victor Hernandez
d79e7e0e12 Map operands of all function-local metadata, not just metadata passed to llvm.dbg.declare intrinsics 
llvm-svn: 93979
 2010-01-20 05:49:59 +00:00
Devang Patel
2cae5754c6 While mapping llvm.dbg.declare intrinsic manually map its operand, if possible, 
because it points to an alloca instruction through metadata.

llvm-svn: 93757
 2010-01-18 19:52:14 +00:00
Bob Wilson
bdd9890c7d Fix a comment typo. 
llvm-svn: 93560
 2010-01-15 21:55:02 +00:00
Chris Lattner
87f86498c3 add a helper function. 
llvm-svn: 93251
 2010-01-12 19:40:54 +00:00
Julien Lerouge
950d1334d1 Fix nondeterministic behavior. 
llvm-svn: 93093
 2010-01-10 01:07:22 +00:00
Julien Lerouge
76c75e82d0 Fix nondeterministic behavior. 
llvm-svn: 93038
 2010-01-09 01:06:49 +00:00
Benjamin Kramer
c233521d45 Convert a ton of simple integer type equality tests to the new predicate. 
llvm-svn: 92760
 2010-01-05 20:07:06 +00:00
Dan Gohman
4b9a39d586 Nick Lewycky pointed out that this code makes changes unconditionally. 
llvm-svn: 92739
 2010-01-05 17:50:58 +00:00
Dan Gohman
e4a22bdd4a Use do+while instead of while for loops which obviously have a 
non-zero trip count. Use SmallVector's pop_back_val().

llvm-svn: 92734
 2010-01-05 16:27:25 +00:00
Dan Gohman
41cc1e1fc4 Make RecursivelyDeleteTriviallyDeadInstructions, 
RecursivelyDeleteDeadPHINode, and DeleteDeadPHIs return a flag
indicating whether they made any changes.

llvm-svn: 92732
 2010-01-05 15:45:31 +00:00
Benjamin Kramer
e90a3c66c4 Avoid going through the LLVMContext for type equality where it's safe to dereference the type pointer. 
llvm-svn: 92726
 2010-01-05 13:12:22 +00:00
David Greene
d0587e9991 Change errs() to dbgs(). 
llvm-svn: 92606
 2010-01-05 01:26:57 +00:00
David Greene
1d478b6184 Change errs() to dbgs(). 
llvm-svn: 92605
 2010-01-05 01:26:54 +00:00
David Greene
e07c8e1668 Change errs() to dbgs(). 
llvm-svn: 92604
 2010-01-05 01:26:52 +00:00
David Greene
991e882855 Change errs() to dbgs(). 
llvm-svn: 92603
 2010-01-05 01:26:49 +00:00
David Greene
94ab26dd76 Change errs() to dbgs(). 
llvm-svn: 92602
 2010-01-05 01:26:45 +00:00
David Greene
2805ab874b Change errs() to dbgs(). 
llvm-svn: 92601
 2010-01-05 01:26:44 +00:00
David Greene
e7608c8372 Change errs() to dbgs(). 
llvm-svn: 92600
 2010-01-05 01:26:41 +00:00