Douglas Gregor [Wed, 22 Jul 2009 00:28:09 +0000 (00:28 +0000)]
Complain if we're entering the context of a dependent nested-name-specifier but
cannot match that nested-name-specifier to a class template or class template
partial specialization.
llvm-svn: 76704
Evan Cheng [Wed, 22 Jul 2009 00:25:27 +0000 (00:25 +0000)]
Let each target determines whether a machine instruction is dead. If true, that allows late codeine passes to delete it.
This is considered a workaround. The problem is some targets are not modeling side effects correctly. PPC is apparently one of those. This patch allows ppc llvm-gcc to bootstrap on Darwin. Once we find out which instruction definitions are wrong, we can remove the PPCInstrInfo workaround.
llvm-svn: 76703
Owen Anderson [Wed, 22 Jul 2009 00:24:57 +0000 (00:24 +0000)]
Get rid of the Pass+Context magic.
llvm-svn: 76702
Chris Lattner [Wed, 22 Jul 2009 00:05:44 +0000 (00:05 +0000)]
reimplement Constant::ContainsRelocations as
Constant::getRelocationInfo(), which has a much simpler
to use API. It still should not be part of libvmcore, but
is better than it was. Also teach it to be smart about
hidden visibility.
llvm-svn: 76700
Dan Gohman [Wed, 22 Jul 2009 00:04:19 +0000 (00:04 +0000)]
Misc. doc fixes following suggestions from Eli.
llvm-svn: 76699
Evan Cheng [Tue, 21 Jul 2009 23:56:01 +0000 (23:56 +0000)]
Fixing cp island pass. Step 1: Determine whether the constant pool offset can be
negative on an individual bases rather than basing on whether it's in thumb
mode.
llvm-svn: 76698
Evan Cheng [Tue, 21 Jul 2009 23:54:22 +0000 (23:54 +0000)]
Fix comment.
llvm-svn: 76693
Douglas Gregor [Tue, 21 Jul 2009 23:53:31 +0000 (23:53 +0000)]
Basic parsing and semantic analysis for out-of-line definitions of the
member functions of class templates, e.g.,
template<typename T>
struct X {
void f(T);
};
template<typename T> X<T>::f(T) { /* ... */ }
llvm-svn: 76692
Mike Stump [Tue, 21 Jul 2009 23:52:50 +0000 (23:52 +0000)]
Revert this, we can now avoid error cascades better.
llvm-svn: 76691
Mike Stump [Tue, 21 Jul 2009 23:50:15 +0000 (23:50 +0000)]
Revert this, was a bug in my new warning code, not the test case.
llvm-svn: 76690
Chris Lattner [Tue, 21 Jul 2009 23:49:55 +0000 (23:49 +0000)]
simplify code now that it is inlined.
llvm-svn: 76689
Lang Hames [Tue, 21 Jul 2009 23:47:33 +0000 (23:47 +0000)]
Exposed PHIElimination pass within CodeGen.
llvm-svn: 76688
Mike Stump [Tue, 21 Jul 2009 23:47:12 +0000 (23:47 +0000)]
Revert this, we have a better way to do this.
llvm-svn: 76687
Chris Lattner [Tue, 21 Jul 2009 23:47:11 +0000 (23:47 +0000)]
Now that RelocBehaviour() is never overloaded, it doesn't need to be
virtual. Just inline it into its two current call sites in preparation
for simplifying the code.
llvm-svn: 76686
Mike Stump [Tue, 21 Jul 2009 23:45:39 +0000 (23:45 +0000)]
Revert this, we have a better way to handle this.
llvm-svn: 76685
Eric Christopher [Tue, 21 Jul 2009 23:43:14 +0000 (23:43 +0000)]
Add intrinsic for __builtin_ia32_vec_set_v4si which turns out to be
pinsrd in sse4.1.
llvm-svn: 76684
Chris Lattner [Tue, 21 Jul 2009 23:42:24 +0000 (23:42 +0000)]
this doesn't break any of the 4 ia64 tests.
llvm-svn: 76683
Chris Lattner [Tue, 21 Jul 2009 23:41:35 +0000 (23:41 +0000)]
alpha doesn't need to redefine this: it only supports PIC codegen anyway.
llvm-svn: 76682
Mike Stump [Tue, 21 Jul 2009 23:41:23 +0000 (23:41 +0000)]
Revert recent change, I now have a better way to solve this (thanks Chris).
llvm-svn: 76681
David Greene [Tue, 21 Jul 2009 23:36:14 +0000 (23:36 +0000)]
Add some support for iterative coalescers to calculate a joined live
range's weight properly. This is turned off right now in the sense that
you'll get an assert if you get into a situation that can only be caused
by an iterative coalescer. All other code paths operate exactly as
before so there is no functional change with this patch. The asserts
should be disabled if/when an iterative coalescer gets added to trunk.
llvm-svn: 76680
Chris Lattner [Tue, 21 Jul 2009 23:36:01 +0000 (23:36 +0000)]
no really, I can spell!
llvm-svn: 76679
Chris Lattner [Tue, 21 Jul 2009 23:34:23 +0000 (23:34 +0000)]
add an API so target-independent codegen can determine if a constant
pool entry will require relocations against it. I implemented this
conservatively for ARM, someone who is knowledgable about it should
see if this can be improved.
llvm-svn: 76678
Ted Kremenek [Tue, 21 Jul 2009 23:27:57 +0000 (23:27 +0000)]
Fix long-lurking bug in ObjCSummaryCache revealed by Torok's recent memory
poisoning changes to DenseMap. We were using an iterator after it had been
invalidated by an insertion into the DenseMap.
llvm-svn: 76677
Dan Gohman [Tue, 21 Jul 2009 23:21:54 +0000 (23:21 +0000)]
Convert instcombine from using using getAnalysis<TargetData> to
getAnalysisIfAvailable<TargetData>.
llvm-svn: 76676
Dan Gohman [Tue, 21 Jul 2009 23:19:40 +0000 (23:19 +0000)]
Permit the IntPtrTy argument to isEliminableCastPair to be null,
to help support use when TargetData is not available.
llvm-svn: 76675
Chris Lattner [Tue, 21 Jul 2009 23:17:26 +0000 (23:17 +0000)]
fix typo
llvm-svn: 76674
Bruno Cardoso Lopes [Tue, 21 Jul 2009 23:13:26 +0000 (23:13 +0000)]
Change ELFCodeEmitter logic to emit the constant pool and jump tables to
their appropriate sections before the code itself. They need to be emitted
before the function because on some targets (x86 but not x86_64) the later
may reference a JT or CP entry address
llvm-svn: 76672
Dan Gohman [Tue, 21 Jul 2009 23:03:19 +0000 (23:03 +0000)]
Replace the original ad-hoc code for determining whether (v pred w) implies
(x pred y) with more thorough code that does more complete canonicalization
before resorting to range checks. This helps it find more cases where
the canonicalized expressions match.
llvm-svn: 76671
Mike Stump [Tue, 21 Jul 2009 22:54:02 +0000 (22:54 +0000)]
Prep for new warning.
llvm-svn: 76670
Evan Cheng [Tue, 21 Jul 2009 22:52:04 +0000 (22:52 +0000)]
Remove a big test case.
llvm-svn: 76669
Chris Lattner [Tue, 21 Jul 2009 22:47:03 +0000 (22:47 +0000)]
Various doc updates from Edward O'Callaghan!
llvm-svn: 76668
Chris Lattner [Tue, 21 Jul 2009 22:42:37 +0000 (22:42 +0000)]
remove the last bits of SectionFlagsForGlobal. There is some flag here that
depends on XS1A, but I think the ReadOnlySection is already set up for this
and there is no testcase that this breaks. If this is really needed, we can
add the appropriate parameterization to TargetAsmInfo in the future to support
this.
llvm-svn: 76667
Chris Lattner [Tue, 21 Jul 2009 22:39:28 +0000 (22:39 +0000)]
don't mask out the small flag and then reapply it later.
llvm-svn: 76666
Chris Lattner [Tue, 21 Jul 2009 22:37:38 +0000 (22:37 +0000)]
if Xcore doesn't support TLS, it doesn't have to worry about thread local LLVM IR, it should be rejected by a front-end.
llvm-svn: 76665
Chris Lattner [Tue, 21 Jul 2009 22:36:53 +0000 (22:36 +0000)]
remove the Xcore implementation of SelectSectionForGlobal. While you have
to twist your brain to see it, I believe it is the same as ELFTargetAsmInfo::SelectSectionForGlobal.
llvm-svn: 76664
Fariborz Jahanian [Tue, 21 Jul 2009 22:36:06 +0000 (22:36 +0000)]
Patch to accomodate Doug's comment on default
destruction of base/members for each destructor AST.
llvm-svn: 76663
Chris Lattner [Tue, 21 Jul 2009 22:32:55 +0000 (22:32 +0000)]
simplify based on the fact that darwin always uses L/l.
llvm-svn: 76662
Chris Lattner [Tue, 21 Jul 2009 22:30:39 +0000 (22:30 +0000)]
make some stuff private.
llvm-svn: 76661
Chris Lattner [Tue, 21 Jul 2009 22:30:25 +0000 (22:30 +0000)]
improve comments.
llvm-svn: 76660
Chris Lattner [Tue, 21 Jul 2009 22:25:52 +0000 (22:25 +0000)]
Remove the XCore custom implementation of MergeableConstSection, relying on
the generic ELF version instead. This will result in its mergable constant
sections getting named ".rodata.cst4" instead of ".cp.const4", but the
linker looks at the section flags, not the name of the section AFAICT.
llvm-svn: 76659
Bob Wilson [Tue, 21 Jul 2009 21:56:46 +0000 (21:56 +0000)]
Fix ocaml tests for 64-bit MacOS systems. LLVM is currently built
as 32-bit code by default, and if gcc defaults to 64-bit code then ocamlc
requires a -cc "gcc -arch i386" option. We were hardcoding -cc g++
and throwing away any other compiler options that were determined when
ocamlc was configured and built.
llvm-svn: 76658
Bob Wilson [Tue, 21 Jul 2009 21:52:57 +0000 (21:52 +0000)]
Reorder the "Metadata" entry to match the C bindings.
This fixes a regression in the vmcore.ml dejagnu test.
llvm-svn: 76657
Ted Kremenek [Tue, 21 Jul 2009 21:48:25 +0000 (21:48 +0000)]
Remove stale comment and fix RUN line.
llvm-svn: 76656
Daniel Dunbar [Tue, 21 Jul 2009 21:33:58 +0000 (21:33 +0000)]
Add missing include
llvm-svn: 76655
Chris Lattner [Tue, 21 Jul 2009 21:29:08 +0000 (21:29 +0000)]
inline a trivial method into its only call site and fix indentation of cases
llvm-svn: 76654
Chris Lattner [Tue, 21 Jul 2009 21:26:32 +0000 (21:26 +0000)]
Remove some overridden functions in XCoreTargetAsmInfo that are
implemented exactly the same way as its ELFTargetAsmInfo subclass
has them.
llvm-svn: 76653
Ted Kremenek [Tue, 21 Jul 2009 21:21:04 +0000 (21:21 +0000)]
Add test case for PR 4596, which is already fixed due to Steve Naroff's overhaul of the Objective-C type system, but isn't in a checker build yet.
llvm-svn: 76648
Ted Kremenek [Tue, 21 Jul 2009 21:16:46 +0000 (21:16 +0000)]
Patch by Stefan Bühler: Escape text in macro expansion when emitting HTML in the HTMLRewriter.
This fixes PR 4602.
llvm-svn: 76647
Chris Lattner [Tue, 21 Jul 2009 21:12:58 +0000 (21:12 +0000)]
revert r76602, 76603, and r76615, pending design discussions.
llvm-svn: 76646
Chris Lattner [Tue, 21 Jul 2009 21:09:35 +0000 (21:09 +0000)]
minor cleanups.
llvm-svn: 76645
Ted Kremenek [Tue, 21 Jul 2009 21:09:31 +0000 (21:09 +0000)]
Update stale comment.
llvm-svn: 76644
Mike Stump [Tue, 21 Jul 2009 21:06:08 +0000 (21:06 +0000)]
Prep for new warning.
llvm-svn: 76643
Ted Kremenek [Tue, 21 Jul 2009 21:03:50 +0000 (21:03 +0000)]
Reorder files.
llvm-svn: 76642
Ted Kremenek [Tue, 21 Jul 2009 21:03:30 +0000 (21:03 +0000)]
Fix PR 4594 by refactoring almost all casting logic from GRExprEngine::VisitCast
to SValuator::EvalCast. In the process, the StoreManagers now use this new cast
machinery, and the hack in GRExprEngine::EvalBind to handle implicit casts
involving OSAtomicCompareAndSwap and friends has been removed (and replaced with
logic closer to the logic specific to those functions).
llvm-svn: 76641
Mike Stump [Tue, 21 Jul 2009 21:03:09 +0000 (21:03 +0000)]
Prep for new warning.
llvm-svn: 76640
Owen Anderson [Tue, 21 Jul 2009 20:55:28 +0000 (20:55 +0000)]
Privatize the ConstantArray table.
llvm-svn: 76639
Mike Stump [Tue, 21 Jul 2009 20:52:43 +0000 (20:52 +0000)]
Prep for new warning.
llvm-svn: 76638
Mike Stump [Tue, 21 Jul 2009 20:50:41 +0000 (20:50 +0000)]
Prep for new warning.
llvm-svn: 76637
Torok Edwin [Tue, 21 Jul 2009 20:27:10 +0000 (20:27 +0000)]
Add a few fairly obvious API changes I noticed while porting some old code.
llvm-svn: 76636
Eli Friedman [Tue, 21 Jul 2009 20:15:24 +0000 (20:15 +0000)]
Missed a piece of the commit to remove the shift flavor.
llvm-svn: 76635
Owen Anderson [Tue, 21 Jul 2009 20:13:12 +0000 (20:13 +0000)]
Privatize the first of the value maps.
llvm-svn: 76634
Eli Friedman [Tue, 21 Jul 2009 20:12:16 +0000 (20:12 +0000)]
Remove shift amount flavor. It isn't actually complete enough to
be useful, and it's currently unused. (Some issues: it isn't actually
rich enough to capture the semantics on many architectures, and
semantics can vary depending on the type being shifted.)
llvm-svn: 76633
Daniel Dunbar [Tue, 21 Jul 2009 20:06:58 +0000 (20:06 +0000)]
driver: Print --version on stdout, to match gcc.
- Patch by Jean-Daniel Dupas
llvm-svn: 76632
Eli Friedman [Tue, 21 Jul 2009 20:05:43 +0000 (20:05 +0000)]
Remove a couple of already-implemented notes.
llvm-svn: 76631
Eli Friedman [Tue, 21 Jul 2009 19:28:10 +0000 (19:28 +0000)]
Misc fixes for -Wreorder:
1. Make it work correctly with anonymous unions.
2. Don't compute it if the warning isn't enabled.
3. Optimize the algorithm slightly to make it linear time in the
case where we don't produce any warnings.
llvm-svn: 76630
Evan Cheng [Tue, 21 Jul 2009 19:25:09 +0000 (19:25 +0000)]
80 col violation.
llvm-svn: 76629
Mike Stump [Tue, 21 Jul 2009 19:03:43 +0000 (19:03 +0000)]
Prep for new warning.
llvm-svn: 76628
Mike Stump [Tue, 21 Jul 2009 19:02:55 +0000 (19:02 +0000)]
Prep for new warning.
llvm-svn: 76627
Mike Stump [Tue, 21 Jul 2009 19:01:48 +0000 (19:01 +0000)]
Prep for new warning.
llvm-svn: 76626
Mike Stump [Tue, 21 Jul 2009 19:01:31 +0000 (19:01 +0000)]
Prep for new warning.
llvm-svn: 76625
Mike Stump [Tue, 21 Jul 2009 19:01:01 +0000 (19:01 +0000)]
Prep for new warning.
llvm-svn: 76624
Mike Stump [Tue, 21 Jul 2009 19:00:12 +0000 (19:00 +0000)]
Prep for new warning.
llvm-svn: 76623
Douglas Gregor [Tue, 21 Jul 2009 18:59:28 +0000 (18:59 +0000)]
Make Sema::ActOnCXXEnterDeclaratorScope robust against failures to compute
the declaration context, as occurs with out-of-line class template member
definitions.
llvm-svn: 76622
Mike Stump [Tue, 21 Jul 2009 18:59:16 +0000 (18:59 +0000)]
Prep for new warning.
llvm-svn: 76621
Mike Stump [Tue, 21 Jul 2009 18:58:50 +0000 (18:58 +0000)]
Prep for new warning.
llvm-svn: 76620
Mike Stump [Tue, 21 Jul 2009 18:58:15 +0000 (18:58 +0000)]
Prep for new warning.
llvm-svn: 76619
Mike Stump [Tue, 21 Jul 2009 18:57:14 +0000 (18:57 +0000)]
Prep for new warning.
llvm-svn: 76617
Mike Stump [Tue, 21 Jul 2009 18:56:34 +0000 (18:56 +0000)]
Prep for new warning.
llvm-svn: 76616
David Greene [Tue, 21 Jul 2009 18:56:32 +0000 (18:56 +0000)]
Prefix IR dumps with LiveInterval indices when possible. This turns
this:
%ESI<def> = MOV32rr %EDI<kill>
ADJCALLSTACKDOWN64 0, %RSP<imp-def>, %EFLAGS<imp-def,dead>, %RSP<imp-use>
%reg1027<def> = MOVZX64rr32 %ESI
%reg1027<def> = ADD64ri8 %reg1027, 15, %EFLAGS<imp-def,dead>
%reg1027<def> = AND64ri8 %reg1027, -16, %EFLAGS<imp-def,dead>
%RDI<def> = MOV64rr %RSP
%RDI<def> = SUB64rr %RDI, %reg1027<kill>, %EFLAGS<imp-def,dead>
%RSP<def> = MOV64rr %RDI
into this:
4 %reg1024<def> = MOV32rr %EDI<kill>
12 ADJCALLSTACKDOWN64 0, %RSP<imp-def>, %EFLAGS<imp-def,dead>, %RSP<imp-use>
20 %reg1025<def> = MOVZX64rr32 %reg1024
28 %reg1026<def> = MOV64rr %reg1025<kill>
36 %reg1026<def> = ADD64ri8 %reg1026, 15, %EFLAGS<imp-def,dead>
44 %reg1027<def> = MOV64rr %reg1026<kill>
52 %reg1027<def> = AND64ri8 %reg1027, -16, %EFLAGS<imp-def,dead>
60 %reg1028<def> = MOV64rr %RSP
68 %reg1029<def> = MOV64rr %reg1028<kill>
76 %reg1029<def> = SUB64rr %reg1029, %reg1027<kill>, %EFLAGS<imp-def,dead>
84 %RSP<def> = MOV64rr %reg1029
This helps greatly when debugging register allocation and coalescing
problems.
llvm-svn: 76615
Mike Stump [Tue, 21 Jul 2009 18:56:04 +0000 (18:56 +0000)]
Prep for new warning.
llvm-svn: 76614
Mike Stump [Tue, 21 Jul 2009 18:54:29 +0000 (18:54 +0000)]
Prep for new warning.
llvm-svn: 76613
Evan Cheng [Tue, 21 Jul 2009 18:54:14 +0000 (18:54 +0000)]
Add fake v7 itineraries for now.
llvm-svn: 76612
Mike Stump [Tue, 21 Jul 2009 18:52:41 +0000 (18:52 +0000)]
Prep for new warning.
llvm-svn: 76611
Mike Stump [Tue, 21 Jul 2009 18:51:31 +0000 (18:51 +0000)]
Prep for new warning.
llvm-svn: 76610
Mike Stump [Tue, 21 Jul 2009 18:48:43 +0000 (18:48 +0000)]
Prep for new warning.
llvm-svn: 76609
Mike Stump [Tue, 21 Jul 2009 18:46:15 +0000 (18:46 +0000)]
Prep for new warning.
llvm-svn: 76608
Mike Stump [Tue, 21 Jul 2009 18:45:53 +0000 (18:45 +0000)]
Prep for new warning.
llvm-svn: 76607
Mike Stump [Tue, 21 Jul 2009 18:45:22 +0000 (18:45 +0000)]
Prep for new warning.
llvm-svn: 76606
Mike Stump [Tue, 21 Jul 2009 18:44:24 +0000 (18:44 +0000)]
Prep for new warnings about control flow falling off the ends of
functions that return a value. I was going to buffer the whole lot
up, but it should be easier to review if I check them in
incrementally. Most of the forth coming changes either add a return
value, or make it impossible to return, or alter the return type.
llvm-svn: 76605
Chris Lattner [Tue, 21 Jul 2009 18:38:57 +0000 (18:38 +0000)]
make AsmPrinter::doFinalization iterate over the global variables
and call PrintGlobalVariable, allowing elimination and simplification
of various targets.
llvm-svn: 76604
David Greene [Tue, 21 Jul 2009 18:27:05 +0000 (18:27 +0000)]
Add PrefixPrinter arguments to the dump routines for MachineFunction and
MachineBasicBlock. We'll use these shortly.
llvm-svn: 76603
David Greene [Tue, 21 Jul 2009 18:21:46 +0000 (18:21 +0000)]
Add a small utility class to configure IR printers. This will allow
printers to do neat and wonderful things when printing debug
information. The ideas is to allow passes to configer printers to emit
pass-specific information when dumping IR.
llvm-svn: 76602
David Greene [Tue, 21 Jul 2009 18:16:24 +0000 (18:16 +0000)]
Make a bunch of changes suggested by Chris and others to improve
efficiency. Fix a few formatting bugs along the way.
llvm-svn: 76601
Evan Cheng [Tue, 21 Jul 2009 18:15:26 +0000 (18:15 +0000)]
Do not select tSXTB / tSXTH in thumb2 mode.
llvm-svn: 76600
Owen Anderson [Tue, 21 Jul 2009 18:06:41 +0000 (18:06 +0000)]
Update for LLVM API change.
llvm-svn: 76599
Owen Anderson [Tue, 21 Jul 2009 18:03:38 +0000 (18:03 +0000)]
Rename getConstantInt{True|False} to get{True|False} at Chris' behest.
llvm-svn: 76598
Torok Edwin [Tue, 21 Jul 2009 18:01:37 +0000 (18:01 +0000)]
Add freed memory poisoning in !NDEBUG mode for DenseMap.
llvm-svn: 76597
Chris Lattner [Tue, 21 Jul 2009 17:59:36 +0000 (17:59 +0000)]
reduce indentation by using an early exit.
llvm-svn: 76596
Ted Kremenek [Tue, 21 Jul 2009 17:43:20 +0000 (17:43 +0000)]
Update CMake files.
llvm-svn: 76595
Ted Kremenek [Tue, 21 Jul 2009 17:41:55 +0000 (17:41 +0000)]
Use string comparison instead of numeric comparison when comparing digests.
llvm-svn: 76594