Commit Graph

115707 Commits

Author SHA1 Message Date
Bill Wendling addcfcac5c Rename the script to indicate that this is for the TEST=simple tests.
llvm-svn: 142764
2011-10-23 20:14:06 +00:00
Bill Wendling e10675a39d Resurrect the 'find regressions for the TEST=nightly tests' script.
llvm-svn: 142763
2011-10-23 20:13:14 +00:00
Chandler Carruth fd7475e906 Now that we have comparison on probabilities, add some static functions
to get important constant branch probabilities and use them for finding
the best branch out of a set of possibilities.

llvm-svn: 142762
2011-10-23 20:10:34 +00:00
Chandler Carruth 446210b616 Remove a commented out line of code that snuck by my auditing.
llvm-svn: 142761
2011-10-23 20:10:30 +00:00
Peter Collingbourne 7d33cd3a2f Fix grammar for C++11 alignment specifiers, and add a few FIXMEs.
llvm-svn: 142760
2011-10-23 20:07:52 +00:00
Peter Collingbourne e0916cd408 Fix typo (C++1 -> C++11).
llvm-svn: 142759
2011-10-23 20:07:42 +00:00
Peter Collingbourne 9f58d7b789 Diagnose unexpanded parameter packs in member initialisers (including
in-class member initialisers).

llvm-svn: 142758
2011-10-23 18:59:44 +00:00
Peter Collingbourne 05156e3231 Clean up duplicated code in Sema::BuildMemInitializer.
llvm-svn: 142757
2011-10-23 18:59:37 +00:00
Peter Collingbourne 8953e7c923 Improve unexpanded parameter pack diagnostics by using UPPC_Initializer here.
llvm-svn: 142756
2011-10-23 18:59:33 +00:00
Peter Collingbourne e109a2ca59 Attach class template attributes to the templated CXXRecordDecl,
instead of silently discarding them.

As a side effect, this improves diagnostics for constexpr class
templates slightly.

llvm-svn: 142755
2011-10-23 17:07:16 +00:00
Benjamin Kramer 1695466fe3 Move Python.h includes out of the headers into the .cpp file where it's actually used.
Python.h includes a ton of macros that can cause weird behavior down the road.

llvm-svn: 142754
2011-10-23 16:49:03 +00:00
Benjamin Kramer d34a329c68 Remove unused include of deprecated header.
llvm-svn: 142753
2011-10-23 16:31:38 +00:00
Benjamin Kramer cc0ed6ba04 Print branch probabilities as percentages.
50% is much more readable than 5.000000e-01.

llvm-svn: 142752
2011-10-23 11:32:54 +00:00
Benjamin Kramer 929f53f65c Add compare operators to BranchProbability and use it to determine if an edge is hot.
llvm-svn: 142751
2011-10-23 11:19:14 +00:00
Tobias Grosser 2ff8723d5d ScopDetection: Allow to limit the scop detection to a single function
-polly-detect-only=<functionname> allows to limit the scop detection to
a single function.

llvm-svn: 142750
2011-10-23 11:17:06 +00:00
Tobias Grosser f3fbcb8919 Fix buildbot link
llvm-svn: 142749
2011-10-23 11:03:06 +00:00
Tobias Grosser 57bc026634 Link to the new LLVM buildbots
llvm-svn: 142748
2011-10-23 11:01:07 +00:00
Tobias Grosser e5b5797529 www: Describe how to load Polly into clang
llvm-svn: 142747
2011-10-23 10:57:53 +00:00
Tobias Grosser 869b9a318d www: Move example into its own section (making room for another one)
llvm-svn: 142746
2011-10-23 10:57:49 +00:00
Tobias Grosser cf1a26838c Add CloogInfo debug output.
llvm-svn: 142745
2011-10-23 10:57:44 +00:00
Tobias Grosser 008442631b Update to isl 0.08
llvm-svn: 142744
2011-10-23 10:39:40 +00:00
Chandler Carruth bd1be4d01c Completely re-write the algorithm behind MachineBlockPlacement based on
discussions with Andy. Fundamentally, the previous algorithm is both
counter productive on several fronts and prioritizing things which
aren't necessarily the most important: static branch prediction.

The new algorithm uses the existing loop CFG structure information to
walk through the CFG itself to layout blocks. It coalesces adjacent
blocks within the loop where the CFG allows based on the most likely
path taken. Finally, it topologically orders the block chains that have
been formed. This allows it to choose a (mostly) topologically valid
ordering which still priorizes fallthrough within the structural
constraints.

As a final twist in the algorithm, it does violate the CFG when it
discovers a "hot" edge, that is an edge that is more than 4x hotter than
the competing edges in the CFG. These are forcibly merged into
a fallthrough chain.

Future transformations that need te be added are rotation of loop exit
conditions to be fallthrough, and better isolation of cold block chains.
I'm also planning on adding statistics to model how well the algorithm
does at laying out blocks based on the probabilities it receives.

The old tests mostly still pass, and I have some new tests to add, but
the nested loops are still behaving very strangely. This almost seems
like working-as-intended as it rotated the exit branch to be
fallthrough, but I'm not convinced this is actually the best layout. It
is well supported by the probabilities for loops we currently get, but
those are pretty broken for nested loops, so this may change later.

llvm-svn: 142743
2011-10-23 09:18:45 +00:00
Benjamin Kramer 42d0cbd76e Reduce dependencies.
llvm-svn: 142742
2011-10-23 08:38:37 +00:00
Craig Topper 980d59832a Add X86 RORX instruction
llvm-svn: 142741
2011-10-23 07:34:00 +00:00
Cameron Zwarich 057fbb1a10 The element insertion code in scalar replacement doesn't handle incorrect
element types, even though the element extraction code does. It is surprising
that this bug has been here for so long. Fixes <rdar://problem/10318778>.

llvm-svn: 142740
2011-10-23 07:02:10 +00:00
Ted Kremenek 142adc492b [analyzer] Remove LocationContext creation methods from AnalysisManager, and change clients to use AnalysisContext instead.
WIP to remove/reduce ExprEngine's usage of AnalysisManager.

llvm-svn: 142739
2011-10-23 02:31:52 +00:00
Craig Topper e94d277db8 Add X86 MULX instruction for disassembler.
llvm-svn: 142738
2011-10-23 00:33:32 +00:00
Craig Topper 7412aa9886 Remove some duplicate specifying of neverHasSideEffects and mayLoad from X86 multiply instructions.
llvm-svn: 142737
2011-10-22 23:13:53 +00:00
Nick Lewycky 52340ac5f8 Oops! Fix test I forgot to submit as part of r142735.
llvm-svn: 142736
2011-10-22 22:07:31 +00:00
Nick Lewycky 32f8051d66 A non-escaping malloc in the entry block is not unlike an alloca. Do dead-store
elimination on them too.

llvm-svn: 142735
2011-10-22 21:59:35 +00:00
Bob Wilson 02d5fe29ab Fix a typo that completely broke the expansion of complex arguments.
llvm-svn: 142734
2011-10-22 21:42:34 +00:00
Richard Smith 2d4063412c Refactor vector constant expression evaluation to return bool like all the other
const expression evaluation subclasses, and remove some APValue copying and
malloc traffic in the process.

llvm-svn: 142733
2011-10-22 21:10:00 +00:00
Howard Hinnant e4383379ae More windows port work by Ruben Van Boxem
llvm-svn: 142732
2011-10-22 20:59:45 +00:00
Nick Lewycky a6674c7fc9 Make SCEV's brute force analysis stronger in two ways. Firstly, we should be
able to constant fold load instructions where the argument is a constant.
Second, we should be able to watch multiple PHI nodes through the loop; this
patch only supports PHIs in loop headers, more can be done here.

With this patch, we now constant evaluate:
  static const int arr[] = {1, 2, 3, 4, 5};
  int test() {
    int sum = 0;
    for (int i = 0; i < 5; ++i) sum += arr[i];
    return sum;
  }

llvm-svn: 142731
2011-10-22 19:58:20 +00:00
Benjamin Kramer e35328c878 Remove obsolete debugging macro, make helper static.
llvm-svn: 142730
2011-10-22 19:16:39 +00:00
Nadav Rotem aa6fab2484 Fix a typo.w
llvm-svn: 142729
2011-10-22 18:44:51 +00:00
Jim Grosbach dfc072d452 Minor updates.
llvm-svn: 142728
2011-10-22 18:17:32 +00:00
Nadav Rotem 9f5ca0ba7d Added my name to CREDITS.TXT
llvm-svn: 142727
2011-10-22 17:51:04 +00:00
Benjamin Kramer 0d6d098841 Move various generated tables into read-only memory, fixing up const correctness along the way.
llvm-svn: 142726
2011-10-22 16:50:00 +00:00
Benjamin Kramer e889bb14f3 Pack struct better.
llvm-svn: 142725
2011-10-22 15:40:28 +00:00
Nadav Rotem e649d66552 Fix pr11193.
SHL inserts zeros from the right, thus even when the original
sign_extend_inreg value was of 1-bit, we need to sra.

llvm-svn: 142724
2011-10-22 12:39:25 +00:00
NAKAMURA Takumi 084e45bc37 test/Driver/debug.c: Mark this as "REQUIRES: shell".
Former RUN line does not make sense on Win32 hosts.
Win32 hosts would not be expected to set appropriate $PWD.

Latter RUN line might be made valid if PathV2::is_absolute("/foo") would be true. Unfortunately, "/foo" should not be treated as absolute path on Win32.

FYI, on mingw32 with MSYS bash (it has 'shell' feature);
Former) $PWD is set as "X:/hogehoge/test/Driver"
Latter) PWD=/foo sets "X:/root/to/msys/foo" to $PWD.

llvm-svn: 142721
2011-10-22 11:28:32 +00:00
NAKAMURA Takumi b5e6fa2949 lib/Driver/Tools.cpp: Use PathV2::is_absolute() to add -fdebug-compilation-dir. pwd[0] might not be '/' on Win32 hosts.
llvm-svn: 142720
2011-10-22 10:25:25 +00:00
Greg Clayton 607ddc571e Bumped versions for lldb-81 and debugserver-146.
llvm-svn: 142718
2011-10-22 03:38:38 +00:00
Greg Clayton f0705c8b90 Added template support when parsing DWARF into types. We can now use STL
classes in the expression parser.

llvm-svn: 142717
2011-10-22 03:33:13 +00:00
Ted Kremenek 33ba99566f Only emit implicit constant conversion truncation warnings in reachable code. Apparently this is what GCC does, and some code depends on this. Fixes <rdar://problem/10321089>.
llvm-svn: 142716
2011-10-22 02:37:33 +00:00
Jason Molenda a98d106d49 Document nexti, show some shorter versions of long lldb
commands.

llvm-svn: 142715
2011-10-22 02:31:42 +00:00
Ted Kremenek 4b4c51c323 Refactor ThreadSafety to use PostOrderCFGView instead of its own copy (of TopologicallySortedCFG).
llvm-svn: 142714
2011-10-22 02:14:27 +00:00
Ted Kremenek 5abde7cd5e Pull TopologicallySortedCFG out of LiveVariables into its own analysis: PostOrderCFGView.
llvm-svn: 142713
2011-10-22 02:14:23 +00:00
Sean Callanan a76eadd8bb Made the expression parser handle persistent variables
correctly even after the process has quit.

llvm-svn: 142712
2011-10-22 01:58:08 +00:00