Commit Graph

39733 Commits

Author SHA1 Message Date
Anton Korobeynikov 59e6d533bd Fix JIT encoding of trap/ud2 instruction
llvm-svn: 46012
2008-01-15 21:40:02 +00:00
Bill Wendling d21d90cf98 Reformatted. It was confusing the other way. No functionality change.
llvm-svn: 46009
2008-01-15 21:16:32 +00:00
Gordon Henriksen 1390ef4163 Fix a memory correctness error noticed by valgrind (harmless in practice).
Thanks to Duncan Sands for noticing it.

llvm-svn: 46007
2008-01-15 20:02:11 +00:00
Duncan Sands 60bd71603b Testcase for gimplify_expr crash caused by an
unexpected placeholder_expr.

llvm-svn: 46006
2008-01-15 19:55:41 +00:00
Steve Naroff 6fcfd0581d - Change Type::isComplexType() to exlude GCC's complex integer extension. In general, we will keep the lowest level Type predicates "pure" (i.e. true to the C99 spec).
- Modify Sema::UsualArithmeticConversions() to work with the new definition of Type::isComplexType().

This is a nice cleanup and also fixes a bug submitted by Eli (which I've added to the test suite).

llvm-svn: 46005
2008-01-15 19:36:10 +00:00
Evan Cheng 4f7437f8af Oops. Forgot to commit this.
llvm-svn: 46003
2008-01-15 07:49:54 +00:00
Evan Cheng eb30bb7d29 Oops. Forgot to commit this.
llvm-svn: 46002
2008-01-15 07:49:36 +00:00
Anton Korobeynikov 6bbbc4cbfa For PR1839: add initial support for __builtin_trap. llvm-gcc part is missed
as well as PPC codegen

llvm-svn: 46001
2008-01-15 07:02:33 +00:00
Chris Lattner 15346fae70 avoid pasting L + "foo" into L"foo".
llvm-svn: 46000
2008-01-15 05:22:14 +00:00
Chris Lattner 877ca77420 avoid token pasting between identifiers and wide strings:
abc+L"foo" -> abc L"foo", not abcL"foo"

llvm-svn: 45999
2008-01-15 05:14:19 +00:00
Chris Lattner 352ab9b4a2 clarify that byval is valid for any pointer argument.
llvm-svn: 45998
2008-01-15 04:34:22 +00:00
Evan Cheng 4d70ba3134 Rename CCIfStruct to CCIfByVal and CCStructAssign to CCPassByVal. Remove unused parameters of CCStructAssign and add size and alignment requirement info.
llvm-svn: 45997
2008-01-15 03:34:58 +00:00
Evan Cheng 48bdfe63e2 Both x86-32 and x86-64 handle byval parameter attributes.
llvm-svn: 45996
2008-01-15 03:15:41 +00:00
Evan Cheng 5b212ea818 ByVal stack slot alignment should be at least as large as pointer ABI alignment.
llvm-svn: 45995
2008-01-15 03:14:05 +00:00
Evan Cheng a9ecdf6927 Revert my last commit. Not needed.
llvm-svn: 45994
2008-01-15 03:10:35 +00:00
Steve Naroff 8115902d6c Add a comment...
llvm-svn: 45993
2008-01-15 02:05:07 +00:00
Steve Naroff 7f988889ef Rework commit r45976, which was incorrect.
- Add Type::isComplexIntegerType(), Type::getAsComplexIntegerType().
- Don't inlude complex types with Type::isIntegerType(), which is too general.
- Use the new predicates in Sema::UsualArithmeticConversions() to recognize/convert the types.

llvm-svn: 45992
2008-01-15 01:41:59 +00:00
Ted Kremenek e584359716 Removed implicit transitions to a "BlockExit" location; we now handle
the end of the block by processing empty blocks (at BlockEntrance) or
when we have just processed the last statement in a block (at PostStmt).

llvm-svn: 45991
2008-01-15 00:24:08 +00:00
Ted Kremenek 619bd1a8a0 Removed mutation of CFGBlock through operator[]. For now the underlying ASTs,
however, can still be manipulated through the returned Stmt*.

llvm-svn: 45990
2008-01-15 00:18:01 +00:00
Steve Naroff d8c7d484af Make sure Sema::GetTypeForDeclarator() deals with abstract declarators when diagnosing incorrect code.
Bug submitted by Eli.

llvm-svn: 45989
2008-01-14 23:33:18 +00:00
Ted Kremenek 3e743661c5 Added prototype implementation of path-sens. analysis core engine.
llvm-svn: 45986
2008-01-14 23:24:37 +00:00
Ted Kremenek 50fd838140 Added operator[] for CFGBlock. This provides random access to the statements
of a block.

llvm-svn: 45984
2008-01-14 23:18:55 +00:00
Chris Lattner 49f980c334 Fix build issue on cygwin, patch by Sam Bishop
llvm-svn: 45977
2008-01-14 21:47:29 +00:00
Steve Naroff c28f46c6b4 Teach Type::isIntegerType() about GCC's __complex__ integer extensions...
Bug submitted by Eli.

llvm-svn: 45976
2008-01-14 21:38:57 +00:00
Steve Naroff 012484d6c8 Fix crasher when redefining functions. Not 100% pleased with this solution, but it is clearly an improvement. Will discuss with Chris later.
llvm-svn: 45975
2008-01-14 20:51:29 +00:00
Duncan Sands 38987c1dcc Simplify CallInst::hasByValArgument using a new method.
llvm-svn: 45974
2008-01-14 19:57:00 +00:00
Duncan Sands b5ca2e9fcb I noticed that the trampoline straightening transformation could
drop attributes on varargs call arguments.  Also, it could generate
invalid IR if the transformed call already had the 'nest' attribute
somewhere (this can never happen for code coming from llvm-gcc,
but it's a theoretical possibility).  Fix both problems.

llvm-svn: 45973
2008-01-14 19:52:09 +00:00
Chris Lattner 994718417a don't create the post-ra scheduler unless it is enabled.
llvm-svn: 45972
2008-01-14 19:00:06 +00:00
Chris Lattner 4272c12571 remove dead #include
llvm-svn: 45971
2008-01-14 18:45:28 +00:00
Ted Kremenek cb994b6a72 Cleaned up comment.
llvm-svn: 45970
2008-01-14 18:42:38 +00:00
Ted Kremenek 94acc6ac9a Cleaned up comment.
llvm-svn: 45969
2008-01-14 18:41:42 +00:00
Ted Kremenek a018a2c0f4 When serializing CompoundLiteralExpr, serialize out the file scope flag before
serializing the subexpression (Init), as this results in a more efficient
encoding in the bitstream.

llvm-svn: 45967
2008-01-14 18:29:39 +00:00
Steve Naroff d32419de0d Record if a compound literal expression is @ file scope. This allows us to implement C99 6.5.2.5p6. This could have been done without modifying the AST (by checking the decl type and passing the info down to isContextExpr), however we concluded this is more desirable.
Bug/patch by Eli Friedman!

llvm-svn: 45966
2008-01-14 18:19:28 +00:00
Ted Kremenek d065a790d9 Removed 'inline' keywords from methods now defined in ExplodedGraph.cpp.
llvm-svn: 45965
2008-01-14 18:11:35 +00:00
Evan Cheng c8dc2f781f Need a space to separate Make options.
llvm-svn: 45964
2008-01-14 17:58:03 +00:00
Ted Kremenek a0a3e9bb07 Change uses of std::cerr/std::cout to llvm::Lcerr/llvm::cout, and remove
#include<iostream>.

Patch provided by Sam Bishop.

llvm-svn: 45962
2008-01-14 16:44:48 +00:00
Steve Naroff 4871fe0b8f Revert r45951, Chris says it violates the C99 spec.
llvm-svn: 45961
2008-01-14 16:10:57 +00:00
Duncan Sands ae8c041b85 This test is now the same as byval-1.ll, so remove it.
llvm-svn: 45960
2008-01-14 14:57:30 +00:00
Duncan Sands 4e079479bb Test that byval cannot be used with pointers to
types with no size.

llvm-svn: 45959
2008-01-14 14:55:05 +00:00
Anders Carlsson a1afcab6f9 _asm is another valid alias for asm.
llvm-svn: 45957
2008-01-14 07:46:40 +00:00
Duncan Sands 1f5340c023 We now allow byval on fairly general pointer types.
llvm-svn: 45956
2008-01-14 06:53:45 +00:00
Chris Lattner 3c43efc9d1 Improve the FP stackifier to decide all on its own whether
an instruction kills a register or not.  This is cheap and 
easy to do now that instructions record this on their flags,
and this eliminates the second pass of LiveVariables from the
x86 backend.  This speeds up a release llc by ~2.5%.

llvm-svn: 45955
2008-01-14 06:41:29 +00:00
Anders Carlsson 1ba25ca171 Add codegen upport for implicit casts to aggregate exprs.
llvm-svn: 45954
2008-01-14 06:28:57 +00:00
Chris Lattner f14ba63620 add a note
llvm-svn: 45953
2008-01-14 06:27:57 +00:00
Chris Lattner fd65291aa3 Fix ASTContext::typesAreCompatible when analyzing a function type with
proto and function type without proto.  It would never call 
'functionTypesAreCompatible' because they have different type classes.

llvm-svn: 45952
2008-01-14 05:45:46 +00:00
Steve Naroff 090353191c Rewrite Expr::isNullPointerConstant() to deal with multiple levels of explicit casts.
Now, isNullPointerConstant() will return true for the following: "(void*)(double*)0"

llvm-svn: 45951
2008-01-14 02:53:34 +00:00
Evan Cheng b9293c51bd Simplify code.
llvm-svn: 45950
2008-01-14 02:38:45 +00:00
Chris Lattner 26fe7ebc03 Fix the miscompilation of MiBench/consumer-lame that was exposed by Evan's
byval work.  This miscompilation is due to the program indexing an array out
of range and us doing a transformation that broke this.

llvm-svn: 45949
2008-01-14 02:09:12 +00:00
Chris Lattner 9a6db18b04 The isNotSuitableForSRA property is now dead, don't compute it.
llvm-svn: 45948
2008-01-14 01:32:52 +00:00
Chris Lattner ab0537258f Change SRAGlobal to not depend on isNotSuitableForSRA, which makes it very
difficult to understand the invariants.

llvm-svn: 45947
2008-01-14 01:31:05 +00:00