d5d95b0b51
[unwind removal] We no longer have 'unwind' instructions being generated, so
...
remove the code that handles them.
llvm-svn: 149901
2012-02-06 21:16:41 +00:00
1fc2eac062
'unwind' is a keyword, not an instruction.
...
llvm-svn: 149898
2012-02-06 21:02:43 +00:00
b7d8f21acd
[unwind removal] Remove the 'unwind' instruction parsing bits.
...
llvm-svn: 149897
2012-02-06 20:50:27 +00:00
07cfb4b696
Introduce helpers to compute the 32-bit varaints and 64-bit variants of
...
some architectures. These are useful for interacting with multiarch or
bi-arch GCC (or GCC-based) toolchains.
llvm-svn: 149895
2012-02-06 20:46:33 +00:00
206dddda15
Test commit; also removes some trailing whitespace
...
llvm-svn: 149887
2012-02-06 19:03:04 +00:00
9ebc0896e1
The 'unwind' instruction is deprecated and will be removed, making this test
...
obsolete.
llvm-svn: 149880
2012-02-06 18:18:47 +00:00
c0449635b3
Update docs describing objective-c property encoding. This includes support for properties that are not backed by an ivar.
...
llvm-svn: 149879
2012-02-06 18:18:25 +00:00
46f28e0e27
Mention that the 'unwind' instruction is now deprecated.
...
llvm-svn: 149876
2012-02-06 17:58:34 +00:00
4488217f73
DebugInfo: Provide a new hook to encode relationship between a property and an ivar.
...
llvm-svn: 149874
2012-02-06 17:49:43 +00:00
c4e342b733
Update ExceptionDemo to use ConstantDataArray.
...
llvm-svn: 149867
2012-02-06 14:09:13 +00:00
2496717052
X86: Don't call malloc for 4 bits. No functionality change.
...
llvm-svn: 149866
2012-02-06 12:06:18 +00:00
baba1aa001
Make helper static.
...
llvm-svn: 149865
2012-02-06 11:28:19 +00:00
ae87d7b4b2
Hexagon: Remove forbidden iostream includes (it introduces static initializers)
...
Reorder includes while at it.
llvm-svn: 149863
2012-02-06 10:19:29 +00:00
239fdf0f61
Split part of EvaluateFunction into a new EvaluateBlock method. No functionality
...
change.
llvm-svn: 149861
2012-02-06 08:24:44 +00:00
accd351e56
Move some llvm_unreachable's from r149849 out of switch statements to satisfy -Wcovered-switch-default
...
llvm-svn: 149860
2012-02-06 08:17:43 +00:00
1f71057747
Add shuffle decoding support for 256-bit pshufd. Merge vpermilp* and pshufd decoding.
...
llvm-svn: 149859
2012-02-06 07:17:51 +00:00
662beed828
fix indentation
...
llvm-svn: 149857
2012-02-06 05:29:32 +00:00
bbb8eb58f9
fix typo
...
llvm-svn: 149856
2012-02-06 05:29:29 +00:00
40d79093d9
SmallVector's construct_range is the same thing as std::uninitialized_fill, no need to reinvent it.
...
llvm-svn: 149851
2012-02-05 22:48:31 +00:00
a8bd26e896
Simplify code. No functionality change.
...
llvm-svn: 149850
2012-02-05 22:14:48 +00:00
c514b5474a
Convert assert(0) to llvm_unreachable
...
llvm-svn: 149849
2012-02-05 22:14:15 +00:00
46e9b441f0
Efficient Constant Uniquing.
...
llvm-svn: 149848
2012-02-05 20:54:10 +00:00
52da72b12a
Teach GlobalOpt to handle atomic accesses to globals.
...
* Most of the transforms come through intact by having each transformed load or
store copy the ordering and synchronization scope of the original.
* The transform that turns a global only accessed in main() into an alloca
(since main is non-recursive) with a store of the initial value uses an
unordered store, since it's guaranteed to be the first thing to happen in main.
(Threads may have started before main (!) but they can't have the address of a
function local before the point in the entry block we insert our code.)
* The heap-SRoA transforms are disabled in the face of atomic operations. This
can probably be improved; it seems odd to have atomic accesses to an alloca
that doesn't have its address taken.
AnalyzeGlobal keeps track of the strongest ordering found in any use of the
global. This is more information than we need right now, but it's cheap to
compute and likely to be useful.
llvm-svn: 149847
2012-02-05 19:56:38 +00:00
613d6d3b43
DefinesPredicate should only look for def operands. Patch by Ludwig Meier.
...
llvm-svn: 149846
2012-02-05 19:55:04 +00:00
bbd1156b95
Clean up some whitespace and comments. No functionality change.
...
llvm-svn: 149845
2012-02-05 19:48:37 +00:00
b97841e981
Fixing a warning in MSVC (this is also a test commit)
...
llvm-svn: 149844
2012-02-05 19:43:39 +00:00
9066fb5c43
Neaten up this method. Check that if there is only one
...
predecessor then it's Src.
llvm-svn: 149843
2012-02-05 19:43:37 +00:00
994c1103c8
Remove dead test: this was already checked and handled a few lines
...
above.
llvm-svn: 149841
2012-02-05 19:30:06 +00:00
be65578460
Testcase for commit 149833 (use of an uninitialized variable noticed
...
by GCC).
llvm-svn: 149840
2012-02-05 19:27:57 +00:00
12efb16b01
Fix a thinko pointed out by Eli and the buildbots.
...
llvm-svn: 149839
2012-02-05 18:56:50 +00:00
4b613497f0
Reduce the number of dom queries made by GVN's conditional propagation
...
logic by half: isOnlyReachableViaThisEdge was trying to be clever and
handle the case of a branch to a basic block which is contained in a
loop. This costs a domtree lookup and is completely useless due to
GVN's position in the pass pipeline: all loops have preheaders at this
point, which means it is enough for isOnlyReachableViaThisEdge to check
that Dst has only one predecessor. (I checked this theoretical argument
by running over the entire nightly testsuite, and indeed it is so!).
llvm-svn: 149838
2012-02-05 18:25:50 +00:00
268903955c
Reduce the number of non-trivial domtree queries by about 1% when
...
compiling sqlite3, by only doing dom queries after the cheap check
rather than interleaved with it.
llvm-svn: 149836
2012-02-05 15:50:43 +00:00
ae22c60f90
Persuade GCC that there is nothing worth warning about here (there isn't).
...
llvm-svn: 149834
2012-02-05 14:20:11 +00:00
efabc2572f
Don't initialize CV in terms of itself! Spotted by GCC.
...
llvm-svn: 149833
2012-02-05 14:16:09 +00:00
bc3f4730c5
Explain to the compiler why TargetAddr is not used uninitialized later.
...
llvm-svn: 149832
2012-02-05 14:14:35 +00:00
4f4546b73a
Add additional documentation to the extract-and-trunc dagcombine optimization.
...
llvm-svn: 149823
2012-02-05 11:39:23 +00:00
93492b8765
Testing vector code without sse doesn't make much sense.
...
Should bring arm and ppc testers back to life (they default to -mcpu=generic)
llvm-svn: 149821
2012-02-05 11:19:39 +00:00
3ea96a701e
Fix typo and broken link
...
llvm-svn: 149820
2012-02-05 11:17:49 +00:00
cf65eff7a5
Add missing paren
...
llvm-svn: 149817
2012-02-05 09:21:25 +00:00
ee4dab5f1f
Convert assert(0) to llvm_unreachable
...
llvm-svn: 149816
2012-02-05 08:31:47 +00:00
ebd90c58e6
Begin fleshing out more convenience predicates in llvm::Triple and
...
convert at least one client over to use them. Subsequent patches both to
LLVM and Clang will try to convert more people over to a common set of
predicates.
This round of predicates is focused on OS-categorization predicates.
llvm-svn: 149815
2012-02-05 08:26:40 +00:00
c4965bce14
Convert assert(0) to llvm_unreachable
...
llvm-svn: 149814
2012-02-05 07:21:30 +00:00
f9c1291fde
Simplify contains tests using 'count'.
...
llvm-svn: 149813
2012-02-05 06:35:36 +00:00
32c48634db
BBVectorize.cpp: Get rid of comparision to bool to fix a warning.
...
llvm-svn: 149810
2012-02-05 05:47:51 +00:00
4ed7278ff4
Convert assert(0) to llvm_unreachable in X86 Target directory.
...
llvm-svn: 149809
2012-02-05 05:38:58 +00:00
83f3bdaa45
Convert some assert(0) in default of switch statements to llvm_unreachable.
...
llvm-svn: 149808
2012-02-05 03:43:23 +00:00
1d471e31ba
Add target specific node for PMULUDQ. Change patterns to use it and custom lower intrinsics to it. Use it instead of intrinsic to handle 64-bit vector multiplies.
...
llvm-svn: 149807
2012-02-05 03:14:49 +00:00
bb8278a10a
Improve the bitcode reader's handling of constant strings to use
...
ConstantDataArray::getString direction, instead of "boxing" each
byte into a ConstantInt and using ConstantArray::get.
llvm-svn: 149805
2012-02-05 02:41:35 +00:00
6987fdb620
Add a test for the miscompilation my recent ConstantDataArray patches introduced, to make sure
...
we don't regress on it in the future.
llvm-svn: 149803
2012-02-05 02:37:36 +00:00
cf9e8f6968
reapply the patches reverted in r149470 that reenable ConstantDataArray,
...
but with a critical fix to the SelectionDAG code that optimizes copies
from strings into immediate stores: the previous code was stopping reading
string data at the first nul. Address this by adding a new argument to
llvm::getConstantStringInfo, preserving the behavior before the patch.
llvm-svn: 149800
2012-02-05 02:29:43 +00:00
Bill Wendling