Nadav Rotem [Mon, 3 Dec 2012 21:06:35 +0000 (21:06 +0000)]
Add initial support for IF-conversion. This patch implements the first 1/3,
which is the legality of the if-conversion transformation. The next step is to
implement the cost-model for the if-converted code as well as the
vectorization itself.
llvm-svn: 169152
Manuel Klimek [Mon, 3 Dec 2012 20:55:42 +0000 (20:55 +0000)]
Fixes a compile warning and crash in the tests.
The necessity of this fix points to a problem with the design
of the addToken during the optimiation phase, which we need to address
in a much more principled way.
llvm-svn: 169151
Tobias Grosser [Mon, 3 Dec 2012 20:47:57 +0000 (20:47 +0000)]
Update CLooG
This cloog versions passes all tests with isl 0.11.
llvm-svn: 169150
Jyotsna Verma [Mon, 3 Dec 2012 20:39:45 +0000 (20:39 +0000)]
Define unsigned const-ext predicates.
llvm-svn: 169149
Jyotsna Verma [Mon, 3 Dec 2012 20:14:38 +0000 (20:14 +0000)]
Removing unnecessary 'else' statement from the predicates defined in HexagonOperards.td.
llvm-svn: 169148
Argyrios Kyrtzidis [Mon, 3 Dec 2012 19:59:23 +0000 (19:59 +0000)]
Eliminate redundant bitwise operations when using a llvm/ADT/PointerUnion.
For comparison, with this code sample:
PointerUnion<int *, char *> Data;
PointerUnion<int *, char *> foo1() {
Data = new int;
return new int;
}
PointerUnion<int *, char *> foo2() {
Data = new char;
return new char;
}
Before this patch we would get:
define i64 @_Z4foo1v() uwtable ssp {
%1 = tail call noalias i8* @_Znwm(i64 4)
%2 = ptrtoint i8* %1 to i64
%3 = load i64* getelementptr inbounds (%"class.llvm::PointerUnion"* @Data, i64 0, i32 0, i32 0), align 8
%4 = and i64 %3, 1
%.masked.i = and i64 %2, -3
%5 = or i64 %4, %.masked.i
store i64 %5, i64* getelementptr inbounds (%"class.llvm::PointerUnion"* @Data, i64 0, i32 0, i32 0), align 8
%6 = tail call noalias i8* @_Znwm(i64 4)
%7 = ptrtoint i8* %6 to i64
%8 = and i64 %7, -3
ret i64 %8
}
define i64 @_Z4foo2v() uwtable ssp {
%1 = tail call noalias i8* @_Znwm(i64 1)
%2 = ptrtoint i8* %1 to i64
%3 = load i64* getelementptr inbounds (%"class.llvm::PointerUnion"* @Data, i64 0, i32 0, i32 0), align 8
%4 = and i64 %3, 1
%5 = or i64 %2, %4
%6 = or i64 %5, 2
store i64 %6, i64* getelementptr inbounds (%"class.llvm::PointerUnion"* @Data, i64 0, i32 0, i32 0), align 8
%7 = tail call noalias i8* @_Znwm(i64 1)
%8 = ptrtoint i8* %7 to i64
%9 = or i64 %8, 2
ret i64 %9
}
After the patch:
define i64 @_Z4foo1v() uwtable ssp {
%1 = tail call noalias i8* @_Znwm(i64 4)
%2 = ptrtoint i8* %1 to i64
store i64 %2, i64* getelementptr inbounds (%"class.llvm::PointerUnion"* @Data, i64 0, i32 0, i32 0), align 8
%3 = tail call noalias i8* @_Znwm(i64 4)
%4 = ptrtoint i8* %3 to i64
ret i64 %4
}
declare noalias i8* @_Znwm(i64)
define i64 @_Z4foo2v() uwtable ssp {
%1 = tail call noalias i8* @_Znwm(i64 1)
%2 = ptrtoint i8* %1 to i64
%3 = or i64 %2, 2
store i64 %3, i64* getelementptr inbounds (%"class.llvm::PointerUnion"* @Data, i64 0, i32 0, i32 0), align 8
%4 = tail call noalias i8* @_Znwm(i64 1)
%5 = ptrtoint i8* %4 to i64
%6 = or i64 %5, 2
ret i64 %6
}
llvm-svn: 169147
Eli Bendersky [Mon, 3 Dec 2012 19:58:12 +0000 (19:58 +0000)]
Simplify this test a bit because DWARF emission/dumping on some platforms
is not yet good enough for more sophistication. The important goal of this
test is to make sure llc doesn't crash on this IR like it used to.
llvm-svn: 169146
Bill Wendling [Mon, 3 Dec 2012 19:44:25 +0000 (19:44 +0000)]
Add 'getInt64Field()' method to get the signed integer instead of unsigned.
llvm-svn: 169145
Alexey Samsonov [Mon, 3 Dec 2012 19:12:58 +0000 (19:12 +0000)]
Add Clang flags -fsanitize-blacklist and -fno-sanitize-blacklist. Make this flag usable for ASan. Blacklisting can be used to disable sanitizer checks for particular file/function/object.
llvm-svn: 169144
Alexey Samsonov [Mon, 3 Dec 2012 19:09:26 +0000 (19:09 +0000)]
ASan: add blacklist file to ASan pass options. Clang patch for this will follow.
llvm-svn: 169143
Eli Bendersky [Mon, 3 Dec 2012 18:45:45 +0000 (18:45 +0000)]
Fix PR12942: Allow two CUs to be generated from the same source file.
Thanks Eric for the review.
llvm-svn: 169142
Kostya Serebryany [Mon, 3 Dec 2012 18:39:21 +0000 (18:39 +0000)]
[asan] two more bits for __sparc__
llvm-svn: 169141
Greg Clayton [Mon, 3 Dec 2012 18:29:55 +0000 (18:29 +0000)]
Fix clang build issues.
llvm-svn: 169140
Benjamin Kramer [Mon, 3 Dec 2012 18:28:52 +0000 (18:28 +0000)]
libclangFormat is not libclangTooling, but it depends on it.
llvm-svn: 169139
Alexey Samsonov [Mon, 3 Dec 2012 18:28:12 +0000 (18:28 +0000)]
Fix PR14474: don't emit debug info for interface types in -gline-tables-only mode.
llvm-svn: 169138
Daniel Jasper [Mon, 3 Dec 2012 18:12:45 +0000 (18:12 +0000)]
Initial version of formatting library.
This formatting library will be used by a stand-alone clang-format tool
and can also be used when writing other refactorings.
Manuel's original design document:
https://docs.google.com/a/google.com/document/d/1gpckL2U_6QuU9YW2L1ABsc4Fcogn5UngKk7fE5dDOoA/edit
The library can already successfully format itself.
Review: http://llvm-reviews.chandlerc.com/D80
llvm-svn: 169137
Greg Clayton [Mon, 3 Dec 2012 17:50:07 +0000 (17:50 +0000)]
<rdar://problem/
12742973>
Forwarding a fix for a crasher in the demangler.
llvm-svn: 169136
Nadav Rotem [Mon, 3 Dec 2012 17:34:44 +0000 (17:34 +0000)]
Teach the jump threading optimization to stop scanning the basic block when calculating the cost after passing the threshold.
llvm-svn: 169135
Jakob Stoklund Olesen [Mon, 3 Dec 2012 17:31:11 +0000 (17:31 +0000)]
Revert r169039, "Aggregate pass execution time report by pass ID instead of pass instance."
llvm-svn: 169134
Chandler Carruth [Mon, 3 Dec 2012 17:02:12 +0000 (17:02 +0000)]
Sort the #include lines for the include/... tree with the script.
AKA: Recompile *ALL* the source code!
This one went much better. No manual edits here. I spot-checked for
silliness and grep-checked for really broken edits and everything seemed
good. It all still compiles. Yell if you see something that looks goofy.
llvm-svn: 169133
Chandler Carruth [Mon, 3 Dec 2012 17:01:46 +0000 (17:01 +0000)]
Teach the include sorter to not choke on include guards. ;]
Kind of important when prepping the include/... tree version of the sort
changes.
llvm-svn: 169132
Chandler Carruth [Mon, 3 Dec 2012 16:50:05 +0000 (16:50 +0000)]
Use the new script to sort the includes of every file under lib.
Sooooo many of these had incorrect or strange main module includes.
I have manually inspected all of these, and fixed the main module
include to be the nearest plausible thing I could find. If you own or
care about any of these source files, I encourage you to take some time
and check that these edits were sensible. I can't have broken anything
(I strictly added headers, and reordered them, never removed), but they
may not be the headers you'd really like to identify as containing the
API being implemented.
Many forward declarations and missing includes were added to a header
files to allow them to parse cleanly when included first. The main
module rule does in fact have its merits. =]
llvm-svn: 169131
Chris Lattner [Mon, 3 Dec 2012 16:08:32 +0000 (16:08 +0000)]
clarify that this isn't lld.
llvm-svn: 169130
Daniel Jasper [Mon, 3 Dec 2012 15:43:25 +0000 (15:43 +0000)]
Make hasDeclaration work for enums.
llvm-svn: 169129
Kostya Serebryany [Mon, 3 Dec 2012 15:00:33 +0000 (15:00 +0000)]
[tsan] add CompactSizeClassMap as an alternative (more compact) size class map. Not used yet.
llvm-svn: 169128
Tobias Grosser [Mon, 3 Dec 2012 14:44:14 +0000 (14:44 +0000)]
Use isl 0.11
We now switch to the newly released isl 0.11. This adds a couple of bug fixes
on top of the recent update.
llvm-svn: 169127
Edwin Vane [Mon, 3 Dec 2012 14:32:13 +0000 (14:32 +0000)]
Test commit: reformat comment.
llvm-svn: 169126
Chandler Carruth [Mon, 3 Dec 2012 14:23:44 +0000 (14:23 +0000)]
Add a completely hack-ish tool to sort includes according to the coding
standards.
I am a terrible Python programmer. Patches more the welcome. Please tell
me how this should look if it should look differently. It's just a tiny
little script so it didn't make sense to go through pre-commit review,
especially as someone who actually knows python may want to just rip it
apart and do it The Right Way.
I will be preparing a commit shortly that uses this script to
canonicalize *all* of the #include lines in LLVM. Really, all of them.
llvm-svn: 169125
Evgeniy Stepanov [Mon, 3 Dec 2012 13:20:43 +0000 (13:20 +0000)]
Add -fsanitize=memory.
llvm-svn: 169124
Kostya Serebryany [Mon, 3 Dec 2012 11:57:01 +0000 (11:57 +0000)]
[tsan] remove unused InternalAllocBlock as part of larger refactoring
llvm-svn: 169123
Dmitry Vyukov [Mon, 3 Dec 2012 11:45:34 +0000 (11:45 +0000)]
tsan: describe global vars (module+offset for now)
llvm-svn: 169122
Kostya Serebryany [Mon, 3 Dec 2012 11:37:03 +0000 (11:37 +0000)]
[asan] EXPECT_EQ => ASSERT_EQ around pthread_create/pthread_join calls
llvm-svn: 169121
Chandler Carruth [Mon, 3 Dec 2012 10:59:55 +0000 (10:59 +0000)]
Remove some buggy and apparantly unnecessary code from SROA.
The partitioning logic attempted to handle uses of an alloca with an
offset starting before the alloca so long as the use had some overlap
with the alloca itself. However, there was a bug where we tested
'(uint64_t)Offset >= AllocSize' without first checking whether 'Offset'
was positive. As a consequence, essentially every negative offset (that
is, starting *before* the alloca does) would be thrown out, even if it
was overlapping. The subsequent code to throw out negative offsets which
were actually non-overlapping was essentially dead. The code to *handle*
overlapping negative offsets was actually dead!
I've just removed all of this, and taught SROA to discard any uses which
start prior to the alloca from the beginning. It has the lovely property
of simplifying the code. =] All the tests still pass, and in fact no new
tests are needed as this is already covered by our testsuite. Fixing the
code so that negative offsets work the way the comments indicate they
were supposed to work causes regressions. That's how I found this.
Anyways, this is all progress in the correct direction -- tightening up
SROA to be maximally aggressive. Some day, I really hope to turn
out-of-bounds accesses to an alloca into 'unreachable'.
llvm-svn: 169120
Nuno Lopes [Mon, 3 Dec 2012 10:15:03 +0000 (10:15 +0000)]
fix stats for added checks
llvm-svn: 169119
Kostya Serebryany [Mon, 3 Dec 2012 09:43:56 +0000 (09:43 +0000)]
[asan] in asan tests, check all return values of pthread_create/pthread_join. Also add the ASAN_AVOID_EXPENSIVE_TESTS macro to guard the test that creates too many threads
llvm-svn: 169118
Jyotsna Verma [Mon, 3 Dec 2012 06:54:50 +0000 (06:54 +0000)]
Define signed const-ext predicates.
llvm-svn: 169117
Bill Wendling [Mon, 3 Dec 2012 06:34:06 +0000 (06:34 +0000)]
Remove 'deplibs' keyword, since it's no longer used.
llvm-svn: 169116
Eli Bendersky [Mon, 3 Dec 2012 04:10:58 +0000 (04:10 +0000)]
Clean up, bring up-to-date and apply consistent formatting.
This document is a long-time pet peeve :-) More fixes to come.
llvm-svn: 169115
Will Dietz [Sun, 2 Dec 2012 19:50:33 +0000 (19:50 +0000)]
[ubsan] Add flag to enable recovery from checks when possible.
llvm-svn: 169114
Will Dietz [Sun, 2 Dec 2012 19:47:29 +0000 (19:47 +0000)]
[ubsan] Refactor handlers to have separate entry points for aborting.
If user specifies aborting after a recoverable failed check is
appropriate, frontend should emit call to the _abort variant.
Test this behavior with newly added -fsanitize-recover flag.
llvm-svn: 169113
Will Dietz [Sun, 2 Dec 2012 18:43:33 +0000 (18:43 +0000)]
[ubsan] Replace "fatal error" with "runtime error".
llvm-svn: 169112
Nadav Rotem [Sun, 2 Dec 2012 17:14:09 +0000 (17:14 +0000)]
Allow merging multiple store sequences on the same chain.
llvm-svn: 169111
Benjamin Kramer [Sun, 2 Dec 2012 16:40:38 +0000 (16:40 +0000)]
Bring vim keyword lists up to date.
llvm-svn: 169110
Eli Bendersky [Sun, 2 Dec 2012 16:02:41 +0000 (16:02 +0000)]
Fix a bug in FileCheck that wouldn't let define variables as follows:
; CHECK: [[VAR:[a-z]]]
The problem was that to find the end of the regex var definition, it was
simplistically looking for the next ]] and finding the incorrect one. A
better approach is to count nesting of brackets (taking escaping into
account). This way the brackets that are part of the regex can be discovered
and skipped properly, and the ]] ending is detected in the right place.
llvm-svn: 169109
Eli Bendersky [Sun, 2 Dec 2012 15:46:02 +0000 (15:46 +0000)]
Fix an invalid regex in the test
llvm-svn: 169108
Chandler Carruth [Sun, 2 Dec 2012 11:53:27 +0000 (11:53 +0000)]
Simplify the coding standards for #include ordering. The ordering is now
trivially achievable with an editor. I'll likely check in a silly python
script to help with this too.
llvm-svn: 169107
Justin Holewinski [Sun, 2 Dec 2012 02:47:48 +0000 (02:47 +0000)]
Unbreak Sphinx build
llvm-svn: 169106
Eli Bendersky [Sat, 1 Dec 2012 22:21:04 +0000 (22:21 +0000)]
Another fix attempt to Makefile.sphinx - copy the PNGs from tutorial/ as well,
and recursive copying is not required for the tutorial/ directory.
llvm-svn: 169105
Eli Bendersky [Sat, 1 Dec 2012 22:11:59 +0000 (22:11 +0000)]
Attempt to fix Makefile.sphinx to not generate errors while trying to copy
files from tutorial/.svn
llvm-svn: 169104
Eli Bendersky [Sat, 1 Dec 2012 22:03:57 +0000 (22:03 +0000)]
Update FileCheck's documentation to mention recently added feature of
matching a variable defined on the same line.
llvm-svn: 169103
Gregory Szorc [Sat, 1 Dec 2012 21:57:30 +0000 (21:57 +0000)]
[python] Add markup option to disassembler
Patch contributed by Wladimir J. van der Laan <laanwj@gmail.com>
llvm-svn: 169102
Eli Bendersky [Sat, 1 Dec 2012 21:54:48 +0000 (21:54 +0000)]
Support referencing variables defined on the same line.
See http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-
20121126/157198.html
and related discussions.
llvm-svn: 169101
Tobias Grosser [Sat, 1 Dec 2012 21:51:10 +0000 (21:51 +0000)]
Update the recommended isl version
Recent changes in isl:
- Allow analysis of loops during code generation
This simplifies the detection of parallel loops.
- Simplify the way costumized ast printers are defined
This enables us to highlight parallel / vector loops in our debug output.
- Compile time improvements for codegen contexts that include parameters
- Various bug fixes
This update also gets us in sync for the isl 0.11 release.
llvm-svn: 169100
Benjamin Kramer [Sat, 1 Dec 2012 20:58:01 +0000 (20:58 +0000)]
Make helper classes anonymous. Make helper functions static instead of private members so the anonymous class doesn't leak out.
No functionality change.
llvm-svn: 169099
Simon Atanasyan [Sat, 1 Dec 2012 18:27:21 +0000 (18:27 +0000)]
[MIPS] Add -mxgot/-mno-xgot command line options
to enable/disable support of GOT larger than 64k.
llvm-svn: 169098
Benjamin Kramer [Sat, 1 Dec 2012 17:54:07 +0000 (17:54 +0000)]
Add raw_ostream include to pacify MSVC.
llvm-svn: 169097
Benjamin Kramer [Sat, 1 Dec 2012 17:22:05 +0000 (17:22 +0000)]
Update unit tests not to rely on transitive includes.
llvm-svn: 169096
Benjamin Kramer [Sat, 1 Dec 2012 17:12:56 +0000 (17:12 +0000)]
Include pruning and general cleanup.
llvm-svn: 169095
Benjamin Kramer [Sat, 1 Dec 2012 16:35:25 +0000 (16:35 +0000)]
Don't include Type.h in DeclarationName.h.
Recursively prune some includes.
llvm-svn: 169094
Benjamin Kramer [Sat, 1 Dec 2012 15:18:03 +0000 (15:18 +0000)]
Make ASTVector.h standalone without including all of ASTContext.
llvm-svn: 169093
Benjamin Kramer [Sat, 1 Dec 2012 15:09:41 +0000 (15:09 +0000)]
Pull the Attr iteration parts out of Attr.h, so including DeclBase.h doesn't pull in all the generated Attr code.
Required to pull some functions out of line, but this shouldn't have a perf impact.
No functionality change.
llvm-svn: 169092
James Molloy [Sat, 1 Dec 2012 14:44:23 +0000 (14:44 +0000)]
Remove bugzilla link.
llvm-svn: 169091
Eli Bendersky [Sat, 1 Dec 2012 13:50:51 +0000 (13:50 +0000)]
Fix a test that was redefining FileCheck variables while referencing old ones.
In preparation for the FileCheck enhancement to support backreferences.
llvm-svn: 169090
Joey Gouly [Sat, 1 Dec 2012 13:07:22 +0000 (13:07 +0000)]
Switch to using -### as mentioned by chandlerc.
llvm-svn: 169089
Joey Gouly [Sat, 1 Dec 2012 12:15:28 +0000 (12:15 +0000)]
Add -emit-llvm to test/Driver/fsanitize.c to stop this failing on the ARM bot. Approved by d0k.
llvm-svn: 169088
Dmitri Gribenko [Sat, 1 Dec 2012 12:13:48 +0000 (12:13 +0000)]
Documentation: convert WritingAnLLVMBackend.html to reST
llvm-svn: 169087
Benjamin Kramer [Sat, 1 Dec 2012 12:08:08 +0000 (12:08 +0000)]
Add .arcconfig to the repository. Useful if someone wants to use phabricator's command line tool.
llvm-svn: 169086
Benjamin Kramer [Sat, 1 Dec 2012 12:07:58 +0000 (12:07 +0000)]
Add .arcconfig to the repository. Useful if someone wants to use phabricator's command line tool.
llvm-svn: 169085
Benjamin Kramer [Sat, 1 Dec 2012 11:53:32 +0000 (11:53 +0000)]
SROA: Avoid struct and array types early to avoid creating an overly large integer type.
Fixes PR14465.
Differential Revision: http://llvm-reviews.chandlerc.com/D148
llvm-svn: 169084
Zhou Sheng [Sat, 1 Dec 2012 10:54:28 +0000 (10:54 +0000)]
Revert previous check in r168581, r169079 as they are still in code review status.
llvm-svn: 169083
Jason Molenda [Sat, 1 Dec 2012 06:13:29 +0000 (06:13 +0000)]
Don't re-add kexts and the kernel each time we get a notification
that kexts were newly added.
The Darwin userland dynamic loader provides lldb with a list of
newly-added or newly-removed binaries but in the kernel case we
only know that something has changed. DynamicLoaderDarwinKernel
really needs to maintain its own persistent list of kexts that
it has been notified about (most importantly, it will not detect
kext unlods) but for now we'll at least avoid re-adding an already
present kext.
<rdar://problem/
12658487>, <rdar://problem/
12658487>
llvm-svn: 169082
Jason Molenda [Sat, 1 Dec 2012 04:50:09 +0000 (04:50 +0000)]
Bump lldb version num to lldb-178, debugserver version to debugserver-198.
llvm-svn: 169081
Jason Molenda [Sat, 1 Dec 2012 04:46:58 +0000 (04:46 +0000)]
Read full 64 bits of kernel address when locating a
64-bit kernel in ProcessGDBRemote.
<rdar://problem/
12657369>
llvm-svn: 169080
Zhou Sheng [Sat, 1 Dec 2012 04:38:53 +0000 (04:38 +0000)]
The patch is to improve the memory footprint of pass GlobalOpt.
Also check in a case to repeat the issue, on which 'opt -globalopt' consumes 1.6GB memory.
The big memory footprint cause is that current GlobalOpt one by one hoists and stores the leaf element constant into the global array, in each iteration, it recreates the global array initializer constant and leave the old initializer alone. This may result in many obsolete constants left.
For example: we have global array @rom = global [16 x i32] zeroinitializer
After the first element value is hoisted and installed: @rom = global [16 x i32] [ 1, 0, 0, ... ]
After the second element value is installed: @rom = global [16 x 32] [ 1, 2, 0, 0, ... ] // here the previous initializer is obsolete
...
When the transform is done, we have 15 obsolete initializers left useless.
llvm-svn: 169079
Michael Han [Sat, 1 Dec 2012 04:35:48 +0000 (04:35 +0000)]
Fix a typo in comment.
llvm-svn: 169078
Richard Smith [Sat, 1 Dec 2012 03:15:03 +0000 (03:15 +0000)]
Make these functions more clearly express the test they are performing. No functionality change.
llvm-svn: 169077
Alexander Potapenko [Sat, 1 Dec 2012 02:39:45 +0000 (02:39 +0000)]
Add caching to the MemoryMappingLayout class on Linux. This is necessary for the cases when a sandbox prevents ASan from reading the mappings
from /proc/self/maps.
The mappings are currently being cached on each access to /proc/self/maps. In the future we'll need to add an API that allows the client to notify ASan about the sandbox.
llvm-svn: 169076
Richard Smith [Sat, 1 Dec 2012 02:35:44 +0000 (02:35 +0000)]
Consistently use 'needsImplicit<special member>' to determine whether we need
an implicit special member, rather than sometimes using '!hasDeclared<special
member>'. No functionality change.
llvm-svn: 169075
NAKAMURA Takumi [Sat, 1 Dec 2012 02:23:45 +0000 (02:23 +0000)]
VMCore/DebugInfo.cpp: DICompileUnit::getSubprograms(): Check numOperands().
2012-11-30-misched-dbg.ll had crashed. Then (MDNode)N was "!{}".
I am not sure it would be ill-formed or not.
llvm-svn: 169074
Daniel Dunbar [Sat, 1 Dec 2012 01:38:48 +0000 (01:38 +0000)]
MC/AsmParser: Avoid unnecessary use of SourceMgr::FindBufferForLoc()
- Each macro instantiation introduces a new buffer, and FindBufferForLoc() is
linear, so previously macro instantiation could be N^2 for some pathological
inputs.
llvm-svn: 169073
Andrew Trick [Sat, 1 Dec 2012 01:22:49 +0000 (01:22 +0000)]
misched: Fix RegisterPressureTracker handling of DebugVals.
Assertion failed: (TopRPTracker.getPos() == RegionBegin && "bad initial Top tracker").
rdar://
12790302.
llvm-svn: 169072
Andrew Trick [Sat, 1 Dec 2012 01:22:44 +0000 (01:22 +0000)]
misched: Fix the DAG builder to handle an undef operand at ExitSU.
Assertion failed: (VNI && "No value to read by operand")
rdar://
12790267.
llvm-svn: 169071
Andrew Trick [Sat, 1 Dec 2012 01:22:41 +0000 (01:22 +0000)]
misched: Fix LiveInterval update to better handle DebugVal.
Assertion failed: (itr != mi2iMap.end() && "Instruction not found in maps.")
rdar://
12777252.
llvm-svn: 169070
Andrew Trick [Sat, 1 Dec 2012 01:22:38 +0000 (01:22 +0000)]
misched: fix RegionBegin when DebugValues get shuffled to the top.
assert (RemainingInstrs == 0 && "Instruction count mismatch!")
rdar://
12776937.
llvm-svn: 169069
Eli Friedman [Sat, 1 Dec 2012 01:15:54 +0000 (01:15 +0000)]
Try to make the source location information for token pastes a bit more consistent.
Fixes a crash printing diagnostics on the gcc testsuite, and also makes
diagnostic range printing print nicer results for token pastes.
llvm-svn: 169068
Jakob Stoklund Olesen [Sat, 1 Dec 2012 01:06:44 +0000 (01:06 +0000)]
Simplify REG_SEQUENCE lowering.
The TwoAddressInstructionPass takes the machine code out of SSA form by
expanding REG_SEQUENCE instructions into copies. It is no longer
necessary to rewrite the registers used by a REG_SEQUENCE instruction
because the new coalescer algorithm can do it now.
REG_SEQUENCE is just converted to a sequence of sub-register copies now.
llvm-svn: 169067
Richard Smith [Sat, 1 Dec 2012 01:02:45 +0000 (01:02 +0000)]
Remove restriction on combining ubsan with asan or tsan. This has worked for a while.
llvm-svn: 169066
Douglas Gregor [Sat, 1 Dec 2012 01:01:09 +0000 (01:01 +0000)]
Fix the determination of whether a capture refers to an enclosing
scope when dealing with nested blocks. Fixes <rdar://problem/
12778708>.
llvm-svn: 169065
Michael Ilseman [Sat, 1 Dec 2012 00:42:27 +0000 (00:42 +0000)]
Update the emacs mode to recognize fadd, fsum, fmul, fdiv, frem, fcmp, icmp
llvm-svn: 169064
Sean Callanan [Sat, 1 Dec 2012 00:09:34 +0000 (00:09 +0000)]
Added support for PtrToInt to the IR
interpreter.
<rdar://problem/
12657742>
llvm-svn: 169063
Sean Callanan [Sat, 1 Dec 2012 00:08:33 +0000 (00:08 +0000)]
Added logging to the code that determines
whether the current frame is in a C++/Objective-C
class or instance method.
llvm-svn: 169062
Eric Christopher [Fri, 30 Nov 2012 23:59:06 +0000 (23:59 +0000)]
Add some first skeleton work for the DWARF5 Fission proposal. Emit
part of the compile unit CU and start separating out information into
the various sections that will be pulled out later.
WIP.
llvm-svn: 169061
Jakob Stoklund Olesen [Fri, 30 Nov 2012 23:53:00 +0000 (23:53 +0000)]
Convert COPY instructions into KILLs if they have implicit defs.
MachineCopyPropagation doesn't understand super-register liveness well
enough to be able to remove implicit defs of super-registers.
This fixes a problem in ARM/2012-01-26-CopyPropKills.ll that is exposed
by an future TwoAddressInstructionPass change. The KILL instructions are
removed before the machine code is emitted.
llvm-svn: 169060
Bill Wendling [Fri, 30 Nov 2012 23:46:56 +0000 (23:46 +0000)]
Don't emit a warning with an input/output parameter. We assume the user knows what they're doing here.
llvm-svn: 169059
Douglas Gregor [Fri, 30 Nov 2012 23:40:49 +0000 (23:40 +0000)]
Fix test for Windows path separators
llvm-svn: 169058
Douglas Gregor [Fri, 30 Nov 2012 23:32:31 +0000 (23:32 +0000)]
Teach the serialized diagnostic writer to clone() itself, sharing
state so that all of the various clones end up rendering their
diagnostics into the same serialized-diagnostics file. This is
important when we actually want failures during module build to be
reported back to the translation unit that tried to import the
not-yet-built or out-of-date module. <rdar://problem/
12565727>
llvm-svn: 169057
Eric Christopher [Fri, 30 Nov 2012 23:20:43 +0000 (23:20 +0000)]
Add support for fission attributes/forms/operations -> string.
llvm-svn: 169056
Michael Ilseman [Fri, 30 Nov 2012 23:18:21 +0000 (23:18 +0000)]
Vim mode updated to recognize fast-math flags
llvm-svn: 169055
Bill Wendling [Fri, 30 Nov 2012 23:18:12 +0000 (23:18 +0000)]
Don't warn if the input size is less than the register size. Also don't warn if
the output size is greater than the register size. No truncation occurs with
those. Reword warning to make it clearer what's the problem is.
llvm-svn: 169054
Michael Ilseman [Fri, 30 Nov 2012 23:14:52 +0000 (23:14 +0000)]
Removed redundancy in wording
llvm-svn: 169053