Anna Zaks [Wed, 10 Apr 2013 22:06:29 +0000 (22:06 +0000)]
[analyzer] Fix a crash in SyntaxCString checker when given a custom strncat.
Fixes PR13476
llvm-svn: 179228
Hal Finkel [Wed, 10 Apr 2013 22:05:25 +0000 (22:05 +0000)]
Manually remove successors in if conversion when CopyAndPredicateBlock is used
In the simple and triangle if-conversion cases, when CopyAndPredicateBlock is
used because the to-be-predicated block has other predecessors, we need to
explicitly remove the old copied block from the successors list. Normally if
conversion relies on TII->AnalyzeBranch combined with BB->CorrectExtraCFGEdges
to cleanup the successors list, but if the predicated block contained an
un-analyzable branch (such as a now-predicated return), then this will fail.
These extra successors were causing a problem on PPC because it was causing
later passes (such as PPCEarlyReturm) to leave dead return-only basic blocks in
the code.
llvm-svn: 179227
Bill Wendling [Wed, 10 Apr 2013 22:03:59 +0000 (22:03 +0000)]
No need to have this return a bool.
llvm-svn: 179226
Jack Carter [Wed, 10 Apr 2013 22:02:32 +0000 (22:02 +0000)]
Mips specific inline asm memory operand modifier test case
These changes are based on commit responses for r179135.
llvm-svn: 179225
Bill Wendling [Wed, 10 Apr 2013 21:56:52 +0000 (21:56 +0000)]
Move info to CREDITS.TXT file.
llvm-svn: 179224
Kay Tiong Khoo [Wed, 10 Apr 2013 21:52:25 +0000 (21:52 +0000)]
fixed xsave, xsaveopt, xrstor mnemonics with intel syntax; added test cases
llvm-svn: 179223
Eric Christopher [Wed, 10 Apr 2013 21:45:07 +0000 (21:45 +0000)]
Revert "Update the version of dwarf we say we're emitting to at least 3."
temporarily while we work on plumbing through some changes to continue
supporting gdb on darwin.
This reverts commit r179122.
llvm-svn: 179222
Anna Zaks [Wed, 10 Apr 2013 21:42:06 +0000 (21:42 +0000)]
[analyzer] When reporting a leak in RetainCount checker due to an early exit from init, step into init.
The heuristic here (proposed by Jordan) is that, usually, if a leak is due to an early exit from init, the allocation site will be
a call to alloc. Note that in other cases init resets self to [super init], which becomes the allocation site of the object.
llvm-svn: 179221
Bill Wendling [Wed, 10 Apr 2013 21:42:06 +0000 (21:42 +0000)]
Track the compact unwind encoding for when we are unable to generate compact unwind information.
Compact unwind has an encoding for when we're not able to generate compact
unwind and must generate an EH frame instead. Track that, but still emit that CU
encoding.
llvm-svn: 179220
Anna Zaks [Wed, 10 Apr 2013 21:42:02 +0000 (21:42 +0000)]
[analyzer] Cleanup leak warnings: do not print the names of variables from other functions.
llvm-svn: 179219
Eric Christopher [Wed, 10 Apr 2013 21:30:40 +0000 (21:30 +0000)]
For split dwarf we should also run the objcopy action if we're
just assembling a file and have passed the option.
llvm-svn: 179218
Chad Rosier [Wed, 10 Apr 2013 21:30:03 +0000 (21:30 +0000)]
Remove unused arguments.
llvm-svn: 179217
Chad Rosier [Wed, 10 Apr 2013 21:26:02 +0000 (21:26 +0000)]
[driver] Add a -Ofast option, which enables -O3, -ffast-math, and
-fstrict-aliasing.
rdar://
13622687
llvm-svn: 179216
Kay Tiong Khoo [Wed, 10 Apr 2013 21:17:58 +0000 (21:17 +0000)]
fixed to disassemble with tab after mnemonic rather than space
llvm-svn: 179215
Reid Kleckner [Wed, 10 Apr 2013 21:10:39 +0000 (21:10 +0000)]
FileCheck-ify some clang grep tests that use double quotes
The escaping interaction between Python and grep doesn't work on my
system. This change fixes the tests for me.
llvm-svn: 179214
Benjamin Kramer [Wed, 10 Apr 2013 20:50:44 +0000 (20:50 +0000)]
Use a real union for IdentifyingPassPtr.
This avoids a nasty const correctness issue (AnalysisIDs are const, Pass* isn't).
llvm-svn: 179213
Bill Wendling [Wed, 10 Apr 2013 20:13:28 +0000 (20:13 +0000)]
Marking myself as release manager.
If anyone objects please let me know.
llvm-svn: 179212
Preston Gurd [Wed, 10 Apr 2013 20:11:59 +0000 (20:11 +0000)]
In the X86 back end, getMemoryOperandNo() returns the offset
into the operand array of the start of the memory reference descriptor.
Additional code in EncodeInstruction provides an additional adjustment.
This patch places that additional code in a separate function,
called getOperandBias, so that any caller of getMemoryOperandNo
can also call getOperandBias.
llvm-svn: 179211
Chad Rosier [Wed, 10 Apr 2013 20:07:47 +0000 (20:07 +0000)]
Tidy up, fix and simplify a few of the SMLocs. Prior to r179109 the Start SMLoc
wasn't always the start of the operand. If there was a symbol reference, then
Start pointed to that token. It's very likely there are other places that need
to be updated.
llvm-svn: 179210
Jyotsna Verma [Wed, 10 Apr 2013 19:53:26 +0000 (19:53 +0000)]
Add object-emission flag for lit tests. This flag is used
to disable following tests for Hexagon that require direct object
generation support.
DebugInfo/dwarf-public-names.ll
DebugInfo/dwarf-version.ll
DebugInfo/member-pointers.ll
DebugInfo/namespace.ll
DebugInfo/two-cus-from-same-file.ll
Fixes bug 15616 - http://llvm.org/bugs/show_bug.cgi?id=15616
llvm-svn: 179209
Howard Hinnant [Wed, 10 Apr 2013 19:44:03 +0000 (19:44 +0000)]
Add capability to demangle invocation functions for ObjC blocks.
llvm-svn: 179208
Nadav Rotem [Wed, 10 Apr 2013 19:41:36 +0000 (19:41 +0000)]
Make the SLP store-merger less paranoid about function calls. We check for function calls when we check if it is safe to sink instructions.
llvm-svn: 179207
Nadav Rotem [Wed, 10 Apr 2013 18:57:27 +0000 (18:57 +0000)]
We require DataLayout for analyzing the size of stores.
llvm-svn: 179206
Chad Rosier [Wed, 10 Apr 2013 18:46:58 +0000 (18:46 +0000)]
Remove unused variable.
llvm-svn: 179205
Dmitri Gribenko [Wed, 10 Apr 2013 18:43:09 +0000 (18:43 +0000)]
-fparse-all-comments: remove redundant check, as suggested by Fariborz Jahanian
llvm-svn: 179204
Hal Finkel [Wed, 10 Apr 2013 18:30:16 +0000 (18:30 +0000)]
PPC: Don't predicate a diamond with two counter decrements
I've not seen this happen in practice, and probably can't until we start
allowing decrement-counter-based conditional branches to be double predicated,
but just in case, don't allow predication of a diamond in which both sides have
ctr-defining branches. Even though the branching behavior of these can be
predicated, the counter-decrementing behavior cannot be.
llvm-svn: 179199
Chad Rosier [Wed, 10 Apr 2013 18:08:17 +0000 (18:08 +0000)]
[ms-inline asm] Move a few test cases from the 32-bit version to the 64-bit
version as lea is only available in 64-bit mode.
llvm-svn: 179190
Chad Rosier [Wed, 10 Apr 2013 17:35:30 +0000 (17:35 +0000)]
Reapply r179115, but use parsePrimaryExpression a little more judiciously.
Test cases that regressed due to r179115, plus a few more, were added in
r179182. Original commit message below:
[ms-inline asm] Use parsePrimaryExpr in lieu of parseExpression if we need to
parse an identifier. Otherwise, parseExpression may parse multiple tokens,
which makes it impossible to properly compute an immediate displacement.
An example of such a case is the source operand (i.e., [Symbol + ImmDisp]) in
the below example:
__asm mov eax, [Symbol + ImmDisp]
Part of rdar://
13611297
llvm-svn: 179187
Michel Danzer [Wed, 10 Apr 2013 17:17:56 +0000 (17:17 +0000)]
R600/SI: Add pattern for AMDGPUurecip
21 more little piglits with radeonsi.
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 179186
Reed Kotler [Wed, 10 Apr 2013 16:58:04 +0000 (16:58 +0000)]
This is for an experimental option -mips-os16. The idea is to compile all
Mips32 code as Mips16 unless it can't be compiled as Mips 16. For now this
would happen as long as floating point instructions are not needed.
Probably it would also make sense to compile as mips32 if atomic operations
are needed too. There may be other cases too.
A module pass prescans the IR and adds the mips16 or nomips16 attribute
to functions depending on the functions needs.
Mips 16 mode can result in a 40% code compression by utililizing 16 bit
encoding of many instructions.
The hope is for this to replace the traditional gcc way of dealing with
Mips16 code using floating point which involves essentially using soft float
but with a library implemented using mips32 floating point. This gcc
method also requires creating stubs so that Mips32 code can interact with
these Mips 16 functions that have floating point needs. My conjecture is
that in reality this traditional gcc method would never win over this
new method.
I will be implementing the traditional gcc method also. Some of it is already
done but I needed to do the stubs to finish the work and those required
this mips16/32 mixed mode capability.
I have more ideas for to make this new method much better and I think the old
method will just live in llvm for anyone that needs the backward compatibility
but I don't for what reason that would be needed.
llvm-svn: 179185
Peter Collingbourne [Wed, 10 Apr 2013 16:52:15 +0000 (16:52 +0000)]
Use a scheme closer to that of GNU as when deciding the type of a
symbol with multiple .type declarations.
Differential Revision: http://llvm-reviews.chandlerc.com/D607
llvm-svn: 179184
Sebastian Pop [Wed, 10 Apr 2013 16:44:08 +0000 (16:44 +0000)]
fix testcase
llvm-svn: 179183
Chad Rosier [Wed, 10 Apr 2013 16:33:34 +0000 (16:33 +0000)]
[ms-inline asm] Add a few test cases that were regressed by r179115. That
commit was reverted in r179120, but I do plan on reapplying with a fix shortly.
Part of rdar://
13611297
llvm-svn: 179182
Dmitri Gribenko [Wed, 10 Apr 2013 16:31:58 +0000 (16:31 +0000)]
Add testcases for -fparse-all-comments
llvm-svn: 179181
Dmitri Gribenko [Wed, 10 Apr 2013 15:35:17 +0000 (15:35 +0000)]
Add an option to parse all comments as documentation comments
Patch by Amin Shali.
llvm-svn: 179180
Rafael Espindola [Wed, 10 Apr 2013 15:33:44 +0000 (15:33 +0000)]
Template MachOObjectFile over endianness too.
llvm-svn: 179179
Rafael Espindola [Wed, 10 Apr 2013 15:18:39 +0000 (15:18 +0000)]
Simplify the templating a bit.
Since we only ever instantiate with a type that is a MachOType instantiation,
we don't need to pass template argument.
llvm-svn: 179178
Alexander Potapenko [Wed, 10 Apr 2013 15:13:00 +0000 (15:13 +0000)]
[ASan] Do not check the shadow of NULL argument in the time() interceptor.
Add a test for time().
llvm-svn: 179177
Rafael Espindola [Wed, 10 Apr 2013 14:57:48 +0000 (14:57 +0000)]
Move two methods out of line.
llvm-svn: 179176
Kostya Serebryany [Wed, 10 Apr 2013 13:59:32 +0000 (13:59 +0000)]
[asan] implement callbacks for unaligned loads/stores
Reviewers: samsonov
Reviewed By: samsonov
CC: samsonov, llvm-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D652
llvm-svn: 179175
Vincent Lejeune [Wed, 10 Apr 2013 13:29:20 +0000 (13:29 +0000)]
R600: Add VTX_READ_* and RAT_WRITE_CACHELESS_* when computing cf addr
llvm-svn: 179174
Reid Kleckner [Wed, 10 Apr 2013 13:11:38 +0000 (13:11 +0000)]
[test] Use lit's shell test runner on Windows
Summary:
I did a local comparison between using bash and using lit's runner, and
more of the suite passes with lit than passes with bash. Most of the
bash failures have to do with /dev/null, which is nonsensical on
Windows, but the lit runner handles it.
The lit shell runner is also much faster than bash, so I would expect
most Windows devs would want it by default.
The behavior can be overridden on any OS by setting
LIT_USE_INTERNAL_SHELL to 0 or 1 in the environment.
Reviewers: chapuni, ddunbar
CC: llvm-commits, timurrrr
Differential Revision: http://llvm-reviews.chandlerc.com/D559
llvm-svn: 179173
Tim Northover [Wed, 10 Apr 2013 12:08:57 +0000 (12:08 +0000)]
Revert "TMP"
This reverts commit
e652085eacbec62e4157d08d3f2f875e6e6d5bb4.
llvm-svn: 179172
Tim Northover [Wed, 10 Apr 2013 12:08:35 +0000 (12:08 +0000)]
ARM: Make "SMC" instructions conditional on new TrustZone architecture feature.
These instructions aren't universally available, but depend on a specific
extension to the normal ARM architecture (rather than, say, v6/v7/...) so a new
feature is appropriate.
This also enables the feature by default on A-class cores which usually have
these extensions, to avoid breaking existing code and act as a sensible
default.
llvm-svn: 179171
Tim Northover [Wed, 10 Apr 2013 12:08:25 +0000 (12:08 +0000)]
TMP
llvm-svn: 179170
Joey Gouly [Wed, 10 Apr 2013 10:37:38 +0000 (10:37 +0000)]
Change CloneFunctionInto to always clone Argument attributes induvidually,
rather than checking if the source and destination have the same number of
arguments and copying the attributes over directly.
llvm-svn: 179169
Manuel Klimek [Wed, 10 Apr 2013 09:52:05 +0000 (09:52 +0000)]
Fixes recovering from errors when parsing braced init lists.
Before we would build huge unwrapped lines which take a long time
to optimze.
llvm-svn: 179168
Daniel Jasper [Wed, 10 Apr 2013 09:49:49 +0000 (09:49 +0000)]
Fix labels with trailing comments and cleanup.
Before:
class A {
public : // test
};
After:
class A {
public: // test
};
Also remove duplicate methods calculating properties of AnnotatedTokens
and make them members of AnnotatedTokens so that they are in a common
place.
llvm-svn: 179167
Christian Konig [Wed, 10 Apr 2013 08:39:16 +0000 (08:39 +0000)]
R600/SI: dynamical figure out the reg class of MIMG
Depending on the number of bits set in the writemask.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
llvm-svn: 179166
Christian Konig [Wed, 10 Apr 2013 08:39:08 +0000 (08:39 +0000)]
R600/SI: adjust writemask to only the used components
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
llvm-svn: 179165
Christian Konig [Wed, 10 Apr 2013 08:39:01 +0000 (08:39 +0000)]
R600/SI: remove image sample writemask
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
llvm-svn: 179164
Tobias Grosser [Wed, 10 Apr 2013 07:42:28 +0000 (07:42 +0000)]
ScopDetect: Allow multiplications of the form <param> * <param>
We handle these by treating this result of the multiplication as an additional
parameter.
llvm-svn: 179163
Hal Finkel [Wed, 10 Apr 2013 07:17:47 +0000 (07:17 +0000)]
Cleanup PPCInstrInfo::DefinesPredicate
Implement suggestions made by Bill Schmidt in post-commit review. Thanks!
llvm-svn: 179162
Alexey Samsonov [Wed, 10 Apr 2013 07:00:25 +0000 (07:00 +0000)]
[ASan] fix a typo in legend in error report
llvm-svn: 179161
Tobias Grosser [Wed, 10 Apr 2013 06:55:45 +0000 (06:55 +0000)]
Update formatting to latest version of clang-format
llvm-svn: 179160
Tobias Grosser [Wed, 10 Apr 2013 06:55:31 +0000 (06:55 +0000)]
Support SCoPs with multiple exit edges
Regions that have multiple exit edges are very common. A simple if condition
yields e.g. such a region:
if
/ \
then else
\ /
after
Region: if -> after
This regions contains the bbs 'if', 'then', 'else', but not 'after'. It has
two exit edges 'then' -> 'after' and 'else' -> 'after'.
Previously we scheduled the RegionSimplify pass to translate such regions into
simple regions. With this patch, we now support them natively.
Contributed-by: Star Tan <tanmx_star@yeah.net>
llvm-svn: 179159
Tobias Grosser [Wed, 10 Apr 2013 06:55:20 +0000 (06:55 +0000)]
Codegen: Replace region exit and entries recursively
During code generation we split the original entry and exit basic blocks
of the scop to make room for the newly generated code. To keep the region tree
up to date, we need to update the region tree. This patch ensures that not only
the region of the scop is updated, but also all child regions that share the
same entry or exit block.
We have now test case here, as the bug is only exposed by the subsequent commit.
The test cases of that commit also cover this bug.
Contributed-by: Star Tan <tanmx_star@yeah.net>
llvm-svn: 179158
Tobias Grosser [Wed, 10 Apr 2013 06:54:49 +0000 (06:54 +0000)]
RegionInfo: Add helpers to replace entry/exit recursively
Contributed by: Star Tan <tanmx_star@yeah.net>
llvm-svn: 179157
Hal Finkel [Wed, 10 Apr 2013 06:42:34 +0000 (06:42 +0000)]
PPC: Prep for if conversion of bctr[l]
This adds in-principle support for if-converting the bctr[l] instructions.
These instructions are used for indirect branching. It seems, however, that the
current if converter will never actually predicate these. To do so, it would
need the ability to hoist a few setup insts. out of the conditionally-executed
block. For example, code like this:
void foo(int a, int (*bar)()) { if (a != 0) bar(); }
becomes:
...
beq 0, .LBB0_2
std 2, 40(1)
mr 12, 4
ld 3, 0(4)
ld 11, 16(4)
ld 2, 8(4)
mtctr 3
bctrl
ld 2, 40(1)
.LBB0_2:
...
and it would be safe to do all of this unconditionally with a predicated
beqctrl instruction.
llvm-svn: 179156
Ted Kremenek [Wed, 10 Apr 2013 06:26:26 +0000 (06:26 +0000)]
Handle "typeof" in Objective-C format string checking. This previously crashed.
Yes, this came from actual code.
Fixes <rdar://problem/
13557053>.
llvm-svn: 179155
Richard Smith [Wed, 10 Apr 2013 06:11:48 +0000 (06:11 +0000)]
Add support for computing the exception specification for an inheriting
constructor. This isn't quite perfect (as usual, we don't handle default
arguments correctly yet, and we don't deal with copy/move constructors for
arguments correctly either, but this will be fixed when we implement core issue
1351.
This completes our support for inheriting constructors.
llvm-svn: 179154
John McCall [Wed, 10 Apr 2013 06:08:21 +0000 (06:08 +0000)]
Don't crash when mangling types defined in ObjC class extensions.
The original test case here was mangling a type name for TBAA,
but we can provoke this in C++11 easily enough.
rdar://
13434937
llvm-svn: 179153
Jason Molenda [Wed, 10 Apr 2013 05:58:57 +0000 (05:58 +0000)]
When ObjectFileMachO::ParseSections() notices that it has a truncated file, zero out the
SectionList so we don't try to do anything with this file. Currently we end up crashing
later in the debug session when we read past the end of the file -- this at least gets us
closer with something like ProcessMachCore printing "error: core file has no sections".
<rdar://problem/
13468295>
llvm-svn: 179152
Richard Smith [Wed, 10 Apr 2013 05:48:59 +0000 (05:48 +0000)]
C++11 inheriting constructors: support for inheriting constructor templates.
llvm-svn: 179151
Douglas Gregor [Wed, 10 Apr 2013 05:18:00 +0000 (05:18 +0000)]
<rdar://problem/
13605348> Don't consider invalid user-defined literal operators during overload resolution.
llvm-svn: 179150
Sebastian Pop [Wed, 10 Apr 2013 04:09:12 +0000 (04:09 +0000)]
fix typos
llvm-svn: 179149
Sebastian Pop [Wed, 10 Apr 2013 04:05:18 +0000 (04:05 +0000)]
scop detection: properly instantiate SCEVs to the place where they are used
Fix inspired from
c2d4a0627e95c34a819b9d4ffb4db62daa78dade.
Given the following code
for (i = 0; i < 10; i++) {
;
}
S: A[i] = 0
When translate the data reference A[i] in statement S using scev, we need to
retrieve the scev of 'i' at the location of 'S'. If we do not do this the
scev that we obtain will be expressed as {0,+,1}_for and will reference loop
iterators that do not surround 'S'. What we really want is the scev to be
instantiated to the value of 'i' after the loop. This value is {10}.
This used to crash in:
int loopDimension = getLoopDepth(Expr->getLoop());
isl_aff *LAff = isl_aff_set_coefficient_si(
isl_aff_zero_on_domain(LocalSpace), isl_dim_in, loopDimension, 1);
(gdb) p Expr->dump()
{8,+,8}<nw><%do.body>
(gdb) p getLoopDepth(Expr->getLoop())
$5 = 0
isl_space *Space = isl_space_set_alloc(Ctx, 0, NbLoopSpaces);
isl_local_space *LocalSpace = isl_local_space_from_space(Space);
As we are trying to create a memory access in a stmt that is outside all loops,
LocalSpace has 0 dimensions:
(gdb) p NbLoopSpaces
$12 = 0
(gdb) p Statement.BB->dump()
if.then: ; preds = %do.end
%0 = load float* %add.ptr, align 4
store float %0, float* %q.1.reg2mem, align 4
br label %if.end.single_exit
and so the scev for %add.ptr should be taken at the place where it is used,
i.e., it should be the value on the last iteration of the do.body loop, and not
"{8,+,8}<nw><%do.body>".
llvm-svn: 179148
Rafael Espindola [Wed, 10 Apr 2013 03:48:25 +0000 (03:48 +0000)]
Template the MachO types over endianness.
For now they are still only used as little endian.
llvm-svn: 179147
Rafael Espindola [Wed, 10 Apr 2013 01:58:26 +0000 (01:58 +0000)]
Include the more specific header.
llvm-svn: 179146
Argyrios Kyrtzidis [Wed, 10 Apr 2013 01:53:50 +0000 (01:53 +0000)]
[frontend] -frewrite-includes: turn implicit module imports into @imports.
rdar://
13610250
llvm-svn: 179145
Argyrios Kyrtzidis [Wed, 10 Apr 2013 01:53:46 +0000 (01:53 +0000)]
[frontend] When preprocessing, turn implicit module imports into @imports.
part of rdar://
13610250
llvm-svn: 179144
Argyrios Kyrtzidis [Wed, 10 Apr 2013 01:53:37 +0000 (01:53 +0000)]
[frontend] -frewrite-includes: if there was no inclusion, don't add lineinfo that indicates return from another file.
llvm-svn: 179143
Michael Ilseman [Wed, 10 Apr 2013 01:41:19 +0000 (01:41 +0000)]
Repeat some #line directive tests for the GNU line marker directive.
llvm-svn: 179142
Evan Cheng [Wed, 10 Apr 2013 01:26:07 +0000 (01:26 +0000)]
__sincosf_stret returns sinf / cosf in bits 0:31 and 32:63 of xmm0, not in
xmm0 / xmm1.
rdar://
13599493
llvm-svn: 179141
Andrew Trick [Wed, 10 Apr 2013 01:06:56 +0000 (01:06 +0000)]
Generalize the PassConfig API and remove addFinalizeRegAlloc().
The target hooks are getting out of hand. What does it mean to run
before or after regalloc anyway? Allowing either Pass* or AnalysisID
pass identification should make it much easier for targets to use the
substitutePass and insertPass APIs, and create less need for badly
named target hooks.
llvm-svn: 179140
Michael Ilseman [Wed, 10 Apr 2013 01:04:18 +0000 (01:04 +0000)]
Improve the diagnostics of the number-reading preprocessor directives.
The GNU line marker directive was sharing code with the #line directive, but some of the warnings/errors were reporting as #line directive diagnostics in both cases.
Previously:
#line 11foo1 ==> "#line directive requires a simple digit sequence"
# 11foo1 ==> "#line directive requires a simple digit sequence"
Now, we get:
#line 11foo1 ==> "#line directive requires a simple digit sequence"
# 11foo1 ==> "GNU line marker directive requires a simple digit sequence"
llvm-svn: 179139
Matt Beaumont-Gay [Wed, 10 Apr 2013 00:47:10 +0000 (00:47 +0000)]
Suppress -Wunused-variable for variables declared in headers, which may in
fact be defined and used in another TU.
Reshuffle some test cases because we suppress -Wunused-variable after we've
emitted an error.
This fixes PR15558.
llvm-svn: 179138
Sean Callanan [Tue, 9 Apr 2013 23:22:08 +0000 (23:22 +0000)]
Don't erroneously put FunctionDecls into CXXRecordDecls
if we didn't want to put in a CXXConstructorDecl. This
prevents malformed classes (i.e., classes with regular C
functions as members) from being generated from type
information (and fixes a crash in the test suite).
<rdar://problem/
13550765>
llvm-svn: 179136
Jack Carter [Tue, 9 Apr 2013 23:19:50 +0000 (23:19 +0000)]
Mips specific inline asm operand modifier 'D'
Modifier 'D' is to use the second word of a double integer.
We had previously implemented the pure register varient of
the modifier and this patch implements the memory reference.
#include "stdio.h"
int b[8] = {0,1,2,3,4,5,6,7};
void main()
{
int i;
// The first word. Notice, no 'D'
{asm (
"lw %0,%1;"
: "=r" (i)
: "m" (*(b+4))
);}
printf("%d\n",i);
// The second word
{asm (
"lw %0,%D1;"
: "=r" (i)
: "m" (*(b+4))
);}
printf("%d\n",i);
}
llvm-svn: 179135
Hal Finkel [Tue, 9 Apr 2013 22:58:37 +0000 (22:58 +0000)]
Allow PPC B and BLR to be if-converted into some predicated forms
This enables us to form predicated branches (which are the same conditional
branches we had before) and also a larger set of predicated returns (including
instructions like bdnzlr which is a conditional return and loop-counter
decrement all in one).
At the moment, if conversion does not capture all possible opportunities. A
simple example is provided in early-ret2.ll, where if conversion forms one
predicated return, and then the PPCEarlyReturn pass picks up the other one. So,
at least for now, we'll keep both mechanisms.
llvm-svn: 179134
Jim Ingham [Tue, 9 Apr 2013 22:34:06 +0000 (22:34 +0000)]
Fix the help for unwind-on-error, it no longer controls what happens when an expression hits a breakpoint.
llvm-svn: 179133
Bob Wilson [Tue, 9 Apr 2013 22:15:51 +0000 (22:15 +0000)]
Fix some comment typos.
llvm-svn: 179132
Richard Trieu [Tue, 9 Apr 2013 22:06:27 +0000 (22:06 +0000)]
Remove outdated run lines from tests.
These run lines originally tested that the fix-its were properly applied.
Originally, the fixits were attached to warnings and were applied by -fixit.
Now, the fixits are attached to notes, so nothing happens. These run lines
still manage to pass since Clang will produce an empty output which gets piped
back to Clang. Then Clang produces no error on an empty input.
llvm-svn: 179131
Sean Callanan [Tue, 9 Apr 2013 21:30:48 +0000 (21:30 +0000)]
Hardening so we won't crash if an Objective-C interface
doesn't have a corresponding type.
<rdar://problem/
13596142>
llvm-svn: 179130
Chad Rosier [Tue, 9 Apr 2013 20:58:48 +0000 (20:58 +0000)]
Cleanup. No functional change intended.
llvm-svn: 179129
Edwin Vane [Tue, 9 Apr 2013 20:51:47 +0000 (20:51 +0000)]
Updating cpp11-migrate transform list in ClangTools docs
Author: Philip Dunstan <phil@phildunstan.com>
llvm-svn: 179128
Edwin Vane [Tue, 9 Apr 2013 20:49:49 +0000 (20:49 +0000)]
Adding the AddOverride transform for cpp11-migrate
This transform adds the override specifier to methods that overrides virtual
methods from a base class that don't already have this specifier.
Author: Philip Dunstan <phil@phildunstan.com>
llvm-svn: 179127
Edwin Vane [Tue, 9 Apr 2013 20:46:36 +0000 (20:46 +0000)]
Adding new AST Matchers isVirtual and isOverride
isVirtual - matches CXXMethodDecl nodes for virtual methods
isOverride - matches CXXMethodDecl nodes for methods that override virtual methods from a base class.
Author: Philip Dunstan <phil@philipdunstan.com>
llvm-svn: 179126
Chad Rosier [Tue, 9 Apr 2013 20:44:09 +0000 (20:44 +0000)]
Cleanup. No functional change intended.
llvm-svn: 179125
Rafael Espindola [Tue, 9 Apr 2013 20:35:08 +0000 (20:35 +0000)]
Remove unused method and default values.
llvm-svn: 179124
Argyrios Kyrtzidis [Tue, 9 Apr 2013 20:29:24 +0000 (20:29 +0000)]
[c-index-test] Enable 'display diagnostics' when using the -test-load functionality.
llvm-svn: 179123
Eric Christopher [Tue, 9 Apr 2013 20:22:47 +0000 (20:22 +0000)]
Update the version of dwarf we say we're emitting to at least 3.
Deals with a dwarf2 -> dwarf3 DW_FORM_ref_addr change.
llvm-svn: 179122
Argyrios Kyrtzidis [Tue, 9 Apr 2013 20:03:03 +0000 (20:03 +0000)]
[libclang] In cxtu::getASTUnit(), check for a null CXTranslationUnit.
llvm-svn: 179121
Chad Rosier [Tue, 9 Apr 2013 19:59:12 +0000 (19:59 +0000)]
Revert r179115 as it looks to have killed the ASan tests.
llvm-svn: 179120
Chandler Carruth [Tue, 9 Apr 2013 19:46:46 +0000 (19:46 +0000)]
Rationalize the formatting of these case labels. Having two sorted
columns is essentially impossible to edit.
llvm-svn: 179119
Reed Kotler [Tue, 9 Apr 2013 19:46:01 +0000 (19:46 +0000)]
This patch enables llvm to switch between compiling for mips32/mips64
and mips16 on a per function basis.
Because this patch is somewhat involved I have provide an overview of the
key pieces of it.
The patch is written so as to not change the behavior of the non mixed
mode. We have tested this a lot but it is something new to switch subtargets
so we don't want any chance of regression in the mainline compiler until
we have more confidence in this.
Mips32/64 are very different from Mip16 as is the case of ARM vs Thumb1.
For that reason there are derived versions of the register info, frame info,
instruction info and instruction selection classes.
Now we register three separate passes for instruction selection.
One which is used to switch subtargets (MipsModuleISelDAGToDAG.cpp) and then
one for each of the current subtargets (Mips16ISelDAGToDAG.cpp and
MipsSEISelDAGToDAG.cpp).
When the ModuleISel pass runs, it determines if there is a need to switch
subtargets and if so, the owning pointers in MipsTargetMachine are
appropriately changed.
When 16Isel or SEIsel is run, they will return immediately without doing
any work if the current subtarget mode does not apply to them.
In addition, MipsAsmPrinter needs to be reset on a function basis.
The pass BasicTargetTransformInfo is substituted with a null pass since the
pass is immutable and really needs to be a function pass for it to be
used with changing subtargets. This will be fixed in a follow on patch.
llvm-svn: 179118
Nadav Rotem [Tue, 9 Apr 2013 19:44:35 +0000 (19:44 +0000)]
Add support for bottom-up SLP vectorization infrastructure.
This commit adds the infrastructure for performing bottom-up SLP vectorization (and other optimizations) on parallel computations.
The infrastructure has three potential users:
1. The loop vectorizer needs to be able to vectorize AOS data structures such as (sum += A[i] + A[i+1]).
2. The BB-vectorizer needs this infrastructure for bottom-up SLP vectorization, because bottom-up vectorization is faster to compute.
3. A loop-roller needs to be able to analyze consecutive chains and roll them into a loop, in order to reduce code size. A loop roller does not need to create vector instructions, and this infrastructure separates the chain analysis from the vectorization.
This patch also includes a simple (100 LOC) bottom up SLP vectorizer that uses the infrastructure, and can vectorize this code:
void SAXPY(int *x, int *y, int a, int i) {
x[i] = a * x[i] + y[i];
x[i+1] = a * x[i+1] + y[i+1];
x[i+2] = a * x[i+2] + y[i+2];
x[i+3] = a * x[i+3] + y[i+3];
}
llvm-svn: 179117
Eric Christopher [Tue, 9 Apr 2013 19:42:12 +0000 (19:42 +0000)]
Make check depend on all.
llvm-svn: 179116
Chad Rosier [Tue, 9 Apr 2013 19:34:59 +0000 (19:34 +0000)]
[ms-inline asm] Use parsePrimaryExpr in lieu of parseExpression if we need to
parse an identifier. Otherwise, parseExpression may parse multiple tokens,
which makes it impossible to properly compute an immediate displacement.
An example of such a case is the source operand (i.e., [Symbol + ImmDisp]) in
the below example:
__asm mov eax, [Symbol + ImmDisp]
The existing test cases exercise this patch.
rdar://
13611297
llvm-svn: 179115
Eric Christopher [Tue, 9 Apr 2013 19:23:15 +0000 (19:23 +0000)]
The .dwo section shouldn't contain the unrelocated values (and
therefore not at all) of the pc or statement list. We also don't
need to emit the compilation dir so save so space and time
and don't bother.
Fix up the testcase accordingly and verify that we don't emit
the attributes or the items that they use.
llvm-svn: 179114