Commit Graph

45737 Commits

Author SHA1 Message Date
Dan Gohman f169f81036 Fix the result type of X86's truncate to i8.
llvm-svn: 53688
2008-07-16 16:20:48 +00:00
Dan Gohman 2714059079 Fix the result type of a VECTOR_SHUFFLE+BIT_CONVERT dagcombine. This
was turned up by some new SelectionDAG assertion checks that I'm
working on.

llvm-svn: 53687
2008-07-16 16:13:58 +00:00
Dan Gohman 494fb26521 Fix the name of BreakCriticalMachineEdge.h's include guard
llvm-svn: 53686
2008-07-16 16:04:07 +00:00
Dan Gohman 52636b2a81 Correct a top-level comment.
llvm-svn: 53685
2008-07-16 16:03:31 +00:00
Duncan Sands 2d28e281e9 Add support for promoting and expanding AssertZext
and AssertSext.  Needed when passing huge integer
parameters with the zeroext or signext attributes.

llvm-svn: 53684
2008-07-16 16:03:07 +00:00
Dan Gohman add5195334 Clarify the comments here, to make slightly more clear the
difference in purpose of TargetInstrInfo and TargetInstrDesc,
which isn't immediately obvious from the name.

llvm-svn: 53683
2008-07-16 16:02:59 +00:00
Dan Gohman b010a08d3a Tidy a doxygen comment.
llvm-svn: 53682
2008-07-16 15:59:56 +00:00
Dan Gohman 26ffe2bea6 Fix a comment to say nonnegative instead of positive.
llvm-svn: 53681
2008-07-16 15:57:10 +00:00
Dan Gohman bf98f68265 Add an assert to check for empty flags for MachineMemOperand.
llvm-svn: 53680
2008-07-16 15:56:42 +00:00
Steve Naroff ce2dca186b RewriteObjC::RewriteObjCTryStmt():Don't synthesize a catch begin if there are 0 catch clauses.
This fixes <rdar://problem/5987211> clang ObjC rewriter: @try / @finally block produces unbalanced output.

llvm-svn: 53679
2008-07-16 15:31:30 +00:00
Steve Naroff b067bbd019 Teach RewriteObjC::RewriteObjCMethodDecl() to deal with pointer to function return types.
This fixes <rdar://problem/6034961> clang ObjC rewriter: rewriting methods with function pointer return values does not work

llvm-svn: 53678
2008-07-16 14:40:40 +00:00
Duncan Sands 5b2c2975b8 Test passing of integer parameters for integers
of all sizes from i1 to i256.  The code is not
always that great, for example (x86)
        movw    %di, %ax
        movw    %ax, i17_s
where the store could be directly from %di.

llvm-svn: 53677
2008-07-16 13:37:36 +00:00
Duncan Sands 5efef246ed Test codegen of loads and stores of all integer
sizes from i1 to i256.  The generated code is
like one huge bug report of things that the DAG
combiner fails to simplify!

llvm-svn: 53676
2008-07-16 13:10:20 +00:00
Matthijs Kooijman f22f34f0cc Add a few cases to instcombine's extractvalue testcase.
llvm-svn: 53675
2008-07-16 12:57:25 +00:00
Matthijs Kooijman c9d4e06f3a Un-XFAIL multdeadretval, since instcombine now properly handles the mess deadargelim leaves behind :-)
llvm-svn: 53674
2008-07-16 12:56:52 +00:00
Matthijs Kooijman c1d7477ed2 Redo InstCombiner::visitExtractValueInst. Instead of using the (complicate)
FindInsertedValue, it now performs a number of simple transformations that
should result in the same effect when applied iteratively.

llvm-svn: 53673
2008-07-16 12:55:45 +00:00
Duncan Sands e766b4230e Reorder methods alphabetically. No functionality change.
While this is not a wonderful organizing principle, it
does make it easy to find routines, and clear where to
insert new ones.

llvm-svn: 53672
2008-07-16 11:41:33 +00:00
Duncan Sands c359055fa9 Turn on LegalizeTypes by default.
llvm-svn: 53671
2008-07-16 11:36:51 +00:00
Matthijs Kooijman b85cc9db19 Don't use ++idx_begin when I actually mean idx_begin + 1, especially since we
also use *idx_begin in the same expression, giving unpredictable results.

This fixes this bug: http://lists.cs.uiuc.edu/pipermail/llvmdev/2008-July/015877.html

llvm-svn: 53670
2008-07-16 10:47:35 +00:00
Duncan Sands a0b1ab3c0b The atomic.cmp.swap promotion logic is wrong: it
simply does the atomic.cmp.swap on the larger type,
which means it blows away whatever is sitting in
the bytes just after the memory location, i.e.
causes a buffer overflow.  This really requires
target specific code, which is why LegalizeTypes
doesn't try to handle this case generically.  The
existing (wrong) code in LegalizeDAG will go away
automatically once the type legalization code is
removed from LegalizeDAG so I'm leaving it there
for the moment.  Meanwhile, don't test for this
feature.

llvm-svn: 53669
2008-07-16 08:09:48 +00:00
Argyrios Kyrtzidis 7da34d050b When checking for name collision between a tag and a previously defined namespace, the collision occured even when the tag was in a different nested scope.
Fix it by taking into account the scope when checking for namespace-tag name collisions.

llvm-svn: 53667
2008-07-16 07:45:46 +00:00
Evan Cheng c97094552c Fix PR2296. Do not transform x86_sse2_storel_dq into a full-width store.
llvm-svn: 53666
2008-07-16 07:28:14 +00:00
Argyrios Kyrtzidis 16c04102ed Add 'this' in the comments of Parser::ParseCastExpression to indicate that it is handled.
llvm-svn: 53665
2008-07-16 07:23:27 +00:00
Ted Kremenek de3725367b Updated latest checker build.
llvm-svn: 53664
2008-07-16 02:39:15 +00:00
Evan Cheng cf06fe476f x86-64 PIC JIT fixes: do not generate the extra load for external GV's.
llvm-svn: 53661
2008-07-16 01:34:02 +00:00
Evan Cheng 68bd1546ab X86-64 PIC jump table values are different from x86-32 cases, they are dest - table base.
llvm-svn: 53660
2008-07-16 01:33:08 +00:00
Ted Kremenek ae543148bc Fix regression introduced by http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20080714/006514.html.
The regression was the casts from integers to pointers where not being handled: they would just return UnknownVal.  This would greatly decrease path-sensitivity.

llvm-svn: 53659
2008-07-16 00:23:49 +00:00
Ted Kremenek 803d9ef9d6 Have scan-build control default analyses.
llvm-svn: 53654
2008-07-15 23:41:32 +00:00
Ted Kremenek f46dddac5b Fix transfer function logic in GRSimpleVals for integer casts: only support casts from integers to integers.
This fixes a crash reported by Anders Carlsson!

llvm-svn: 53649
2008-07-15 23:17:54 +00:00
Ted Kremenek 3be3401303 Add -DIBOutlet=__attribute__((iboutlet)) to analyzer arguments.
llvm-svn: 53648
2008-07-15 23:09:14 +00:00
Ted Kremenek ed1f72b67c For the MissingDealloc check, don't treat IBOutlet ivars as being needed to be released
llvm-svn: 53647
2008-07-15 23:04:27 +00:00
Ted Kremenek 72da4ff4f3 Make iboutlet diagnostic lowercase
llvm-svn: 53646
2008-07-15 22:39:36 +00:00
Ted Kremenek 10c8d7653b Use lowercase of attribute iboutlet to not conflict with the macro IBOutlet
llvm-svn: 53645
2008-07-15 22:38:34 +00:00
Ted Kremenek 8e3704d35a Added parsing/sema support for __attribute__ ((IBOutlet)), a clang-specific attribute that the static analyzer will use to recognize what ivars are IBOutlets.
llvm-svn: 53644
2008-07-15 22:26:48 +00:00
Ted Kremenek 3879c5f6e4 Added --status-bugs option to scan-build. By default, the exit status of
scan-build is the same as the exit status of the executed build command. With
this option, the exit status of scan-build is 1 if the analyzer flagged any
bugs, and 0 otherwise.

This addresses: <rdar://problem/6075320>

llvm-svn: 53642
2008-07-15 22:03:09 +00:00
Ted Kremenek 48ea92965b Per Sam Bishop's excellent suggestion, use "system" instead of backticks to invoke sub-commands used by scan-build. This avoids meta-character translation issues caused by a shell subprocess.
llvm-svn: 53640
2008-07-15 20:18:21 +00:00
Dan Gohman efa4160fcc TargetAsmInfo::SectionForGlobal showed up in a profile. Simplify it a little.
llvm-svn: 53639
2008-07-15 18:37:51 +00:00
Dan Gohman 1e5aa12b7d SelectionDAG::AssignNodeIds is unused.
llvm-svn: 53636
2008-07-15 18:29:32 +00:00
Ted Kremenek c7cd410657 Quote file names to better handle paths with spaces.
llvm-svn: 53635
2008-07-15 18:28:16 +00:00
Dan Gohman 1d093846b5 Don't sort SDNodes by their addresses in SelectionDAG::dump. Instead,
just use the AllNodes order, which is at least relatively stable
across runs.

llvm-svn: 53632
2008-07-15 18:18:54 +00:00
Ted Kremenek 28762a5e49 Incremented latest checker build.
llvm-svn: 53631
2008-07-15 18:15:51 +00:00
Ted Kremenek 81bfc074c9 Distinguish between dead stores and dead initializations.
llvm-svn: 53628
2008-07-15 18:06:32 +00:00
Owen Anderson 24768e3dc4 Revert this, as it seems to still be broken.
llvm-svn: 53627
2008-07-15 17:59:02 +00:00
Ted Kremenek 176e2c519d isRetain() and isRelease() now only returns true if "Retain"/"Release" appears in the suffix of a function's name.
llvm-svn: 53621
2008-07-15 17:43:41 +00:00
Ted Kremenek 78e9c0d697 scan-build now propagates up the exit status of the build command.
llvm-svn: 53620
2008-07-15 17:09:28 +00:00
Ted Kremenek e5165722ed Do not enable -warn-objc-missing-dealloc by default.
llvm-svn: 53619
2008-07-15 17:06:44 +00:00
Ted Kremenek 3a787bf977 scan-build now interrogates clang for a list of available analyses, and presents
these as options to the user of scan-build.

llvm-svn: 53618
2008-07-15 17:06:13 +00:00
Ted Kremenek fa89e2f09c Support retain/release tracking for CoreGraphics (CGxxxRef) objects.
llvm-svn: 53617
2008-07-15 16:50:12 +00:00
Owen Anderson 9d1f497a28 Enable local PRE by default.
llvm-svn: 53616
2008-07-15 16:28:23 +00:00
Owen Anderson 53d546e40b Have GVN do a pre-pass over the CFG that folds away unconditional branches where possible. This allows local PRE to be more aggressive.
llvm-svn: 53615
2008-07-15 16:28:06 +00:00