Jordan Rose [Tue, 26 Feb 2013 01:21:27 +0000 (01:21 +0000)]
[analyzer] Don't look through casts when creating pointer temporaries.
Normally, we need to look through derived-to-base casts when creating
temporary object regions (added in r175854). However, if the temporary
is a pointer (rather than a struct/class instance), we need to /preserve/
the base casts that have been applied.
This also ensures that we really do create a new temporary region when
we need to: MaterializeTemporaryExpr and lvalue CXXDefaultArgExprs.
Fixes PR15342, although the test case doesn't include the crash because
I couldn't isolate it.
llvm-svn: 176069
Jordan Rose [Tue, 26 Feb 2013 01:21:21 +0000 (01:21 +0000)]
[analyzer] StackAddrEscapeChecker: strip qualifiers from temporary types.
With the new support for trivial copy constructors, we are not always
consistent about whether a CXXTempObjectRegion gets reused or created
from scratch, which affects whether qualifiers are preserved. However,
we probably don't care anyway.
This also switches to using the current PrintingPolicy for the type,
which means C++ types don't get a spurious 'struct' prefix anymore.
llvm-svn: 176068
Sean Callanan [Tue, 26 Feb 2013 01:12:25 +0000 (01:12 +0000)]
Fixed several problems with class uniq'ing in the
SymbolFileDWARF code:
- If a class is being uniqued to another copy of itself
and the method lists don't match exactly, take a slow
path and at least unique the methods that they have
in common.
- Sort name_to_die maps before querying them. This
would otherwise result in uniquing failures because
looking up a name in a map that contains it would
often fail.
- Tolerate classes in other symbol files in the case
of debugging with .o files rather than with a
.dSYM. We used to assume that the classes being
uniqued were in the same symbol file, causing
unpredictable results.
This will dramatically reduce the number of cases where
a function does not have a valid DeclContext.
<rdar://problem/
12153915>
llvm-svn: 176067
Chad Rosier [Tue, 26 Feb 2013 01:05:31 +0000 (01:05 +0000)]
[fast-isel] Make sure the FastLowerArguments function checks to make sure the
arguments type is a simple type.
rdar://
13290455
llvm-svn: 176066
Enrico Granata [Tue, 26 Feb 2013 00:28:02 +0000 (00:28 +0000)]
Removing leftover printf() statement
llvm-svn: 176065
Jason Molenda [Tue, 26 Feb 2013 00:26:47 +0000 (00:26 +0000)]
If the user has disabled kext loading with the
plugin.dynamic-loader.darwin-kernel.load-kexts setting, don't print
any messages about loading the kexts (which isn't being done) and
don't read the Mach-O headers out of memory (which can be slow and
they're not being used for anything at this point).
llvm-svn: 176064
Greg Clayton [Tue, 26 Feb 2013 00:21:38 +0000 (00:21 +0000)]
<rdar://problem/
13289157>
Set the exception breakpoints more efficiently by specifying two module basenames as module filters for Apple vendor targets.
llvm-svn: 176063
Han Ming Ong [Tue, 26 Feb 2013 00:04:29 +0000 (00:04 +0000)]
<rdar://problem/
13282582>
Really don't call sysctl again when we already have the result.
llvm-svn: 176062
Jim Ingham [Mon, 25 Feb 2013 23:51:06 +0000 (23:51 +0000)]
Fix the .categories, it had "dataformatter" not "dataformatters".
Remove the getCategory from TestDataFormatterObjC.py, since it was superceded by the .categories file,
and didn't work anyway (getCategories currently has to be a method on the test class, not on the test.)
Add a "basic_process" category, and start to find some tests for simple process running sniff tests.
llvm-svn: 176061
Michael Liao [Mon, 25 Feb 2013 23:16:36 +0000 (23:16 +0000)]
Refine fix to PR10499, no functionality change
- Put expensive checking after simple one
llvm-svn: 176060
Enrico Granata [Mon, 25 Feb 2013 23:01:08 +0000 (23:01 +0000)]
Being explicit about how ignoring optparse's exceptions is not a best principle
llvm-svn: 176059
Michael Liao [Mon, 25 Feb 2013 23:01:03 +0000 (23:01 +0000)]
Fix PR10499
- Check whether SSE is available before lowering all 1s vector building with
PCMPEQD, which is only available from SSE2
llvm-svn: 176058
Andrew Kaylor [Mon, 25 Feb 2013 23:00:19 +0000 (23:00 +0000)]
Provide workaround for PR 15130.
This changes the RecordingMemoryManager in lli to use mapped memory rather than malloc to allocate memory for sections and uses a 'near' MemoryBlock to keep the allocations together. This works around a problem in MCJIT where relocations are applied to a generated image immediately oupon generation, which isn't appropriate for the remote case.
llvm-svn: 176057
Will Dietz [Mon, 25 Feb 2013 22:37:49 +0000 (22:37 +0000)]
[ubsan] Emit single check for left shift.
Avoids warning twice on same shift.
llvm-svn: 176056
Chad Rosier [Mon, 25 Feb 2013 22:20:00 +0000 (22:20 +0000)]
Fix wording.
llvm-svn: 176055
Anna Zaks [Mon, 25 Feb 2013 22:10:34 +0000 (22:10 +0000)]
[analyzer] Restrict ObjC type inference to methods that have related result type.
This addresses a case when we inline a wrong method due to incorrect
dynamic type inference. Specifically, when user code contains a method from init
family, which creates an instance of another class.
Use hasRelatedResultType() to find out if our inference rules should be triggered.
llvm-svn: 176054
Chad Rosier [Mon, 25 Feb 2013 22:06:05 +0000 (22:06 +0000)]
Remove extraneous attribute number.
llvm-svn: 176053
Chad Rosier [Mon, 25 Feb 2013 21:59:35 +0000 (21:59 +0000)]
[fast-isel] Add X86FastIsel::FastLowerArguments to handle functions with 6 or
fewer scalar integer (i32 or i64) arguments. It completely eliminates the need
for SDISel for trivial functions.
Also, add the new llc -fast-isel-abort-args option, which is similar to
-fast-isel-abort option, but for formal argument lowering.
llvm-svn: 176052
Greg Clayton [Mon, 25 Feb 2013 21:53:07 +0000 (21:53 +0000)]
<rdar://problem/
13281528>
Fixed issues with the SBModule "sections" property, and with the SBBlock "ranges" attributes.
llvm-svn: 176051
Ted Kremenek [Mon, 25 Feb 2013 21:32:40 +0000 (21:32 +0000)]
[analyzer] Recover all PreStmtPurgeDeadSymbols nodes with a single successor or predecessor.
These nodes are never consulted by any analyzer client code, so they are
used only for machinery for removing dead bindings. Once successor nodes
are generated they can be safely removed.
This greatly reduces the amount of nodes that are generated in some case,
lowering the memory regression when analyzing Sema.cpp introduced by
r176010 from 14% to 2%.
llvm-svn: 176050
Greg Clayton [Mon, 25 Feb 2013 21:20:59 +0000 (21:20 +0000)]
Added missing lldb module namespace.
llvm-svn: 176049
Daniel Malea [Mon, 25 Feb 2013 20:54:19 +0000 (20:54 +0000)]
Skip test that fails intermittently with Clang/GCC on Linux
- was causing buildbot failures due to unexpected pass
llvm-svn: 176048
Edwin Vane [Mon, 25 Feb 2013 20:43:32 +0000 (20:43 +0000)]
Various additions to ASTMatcher library:
New type matchers:
* recordType
* elaboratedType
New narrowing matchers:
* hasQualifier
* namesType
* hasDeclContext
Added tests and updated LibASTMatchersReference.
Reviewers: klimek
llvm-svn: 176047
Edwin Vane [Mon, 25 Feb 2013 20:37:03 +0000 (20:37 +0000)]
Add user documentation for cpp11-migrate
Adding an RST document for cpp11-migrate. This user documentation explains
command line options, transformations, risk level and how it applies to
loop convert, and code examples of transformations.
There is a TODO task under "Risk" for "Loop Convert" to find code examples
that produce incorrect transformations that change semantics. The definition
of risk in loop convert and instances where the confidence level is lowered
will need to be looked at more carefully.
Information for all new features (including verbose output, auto transform)
will be added in a later change.
Author: Jack Yang <jack.wang@intel.com>
Reviewer: grigozavr
llvm-svn: 176046
Michael J. Spencer [Mon, 25 Feb 2013 20:03:09 +0000 (20:03 +0000)]
[ELF][Writer] Fill in dynamic table entries.
llvm-svn: 176045
Anna Zaks [Mon, 25 Feb 2013 19:51:03 +0000 (19:51 +0000)]
Revert "Add more attributes from the command line to functions."
This reverts commit 176009.
The commit is a likely cause of several buildbot failures.
llvm-svn: 176044
Anna Zaks [Mon, 25 Feb 2013 19:50:50 +0000 (19:50 +0000)]
[analyzer] Address Jordan's code review of r175857.
llvm-svn: 176043
Jordan Rose [Mon, 25 Feb 2013 19:45:34 +0000 (19:45 +0000)]
[analyzer] Handle reference parameters with default values.
r175026 added support for default values, but didn't take reference
parameters into account, which expect the default argument to be an
lvalue. Use createTemporaryRegionIfNeeded if we can evaluate the default
expr as an rvalue but the expected result is an lvalue.
Fixes the most recent report of PR12915. The original report predates
default argument support, so that can't be it.
llvm-svn: 176042
Enrico Granata [Mon, 25 Feb 2013 19:44:39 +0000 (19:44 +0000)]
Optimization of the code-running NSDictionary formatter to skip runtime symbol fetching
llvm-svn: 176041
Greg Clayton [Mon, 25 Feb 2013 19:34:57 +0000 (19:34 +0000)]
<rdar://problem/
13286937>
Make sure to not look in self.images when we have a symbolicator with a live process.
llvm-svn: 176040
Jim Ingham [Mon, 25 Feb 2013 19:31:37 +0000 (19:31 +0000)]
Add a log line when debugserver exits, and clean up some of the other standard logs output to make it more useful.
llvm-svn: 176039
Chad Rosier [Mon, 25 Feb 2013 19:13:23 +0000 (19:13 +0000)]
Add a front-end test case for r176036.
llvm-svn: 176038
Andrew Trick [Mon, 25 Feb 2013 19:11:48 +0000 (19:11 +0000)]
pre-RA-sched fix: only reevaluate physreg interferences when necessary.
Fixes rdar:
13279013: scheduler was blowing up on select instructions.
llvm-svn: 176037
Chad Rosier [Mon, 25 Feb 2013 19:06:27 +0000 (19:06 +0000)]
[ms-inline asm] Add support for the pushad/popad mnemonics.
rdar://
13254235
llvm-svn: 176036
Michael Han [Mon, 25 Feb 2013 18:40:11 +0000 (18:40 +0000)]
Fix empty declaration printing.
Don't print the semicolon when visiting an empty declaration because the semicolon
will be printed as a terminator later.
llvm-svn: 176035
Jordan Rose [Mon, 25 Feb 2013 18:36:15 +0000 (18:36 +0000)]
[analyzer] Base regions may be invalid when layered on symbolic regions.
While RegionStore checks to make sure casts on TypedValueRegions are valid,
it does not do the same for SymbolicRegions, which do not have perfect type
info anyway. Additionally, MemRegion::getAsOffset does not take a
ProgramState, so it can't use dynamic type info to determine a better type
for the regions. (This could also be dangerous if the type of a super-region
changes!)
Account for this by checking that a base object region is valid on top of a
symbolic region, and falling back to "symbolic offset" mode if not.
Fixes PR15345.
llvm-svn: 176034
Renato Golin [Mon, 25 Feb 2013 18:25:58 +0000 (18:25 +0000)]
ARM build docs easier for copy&paste
llvm-svn: 176033
Eric Christopher [Mon, 25 Feb 2013 18:19:32 +0000 (18:19 +0000)]
Add forgotten testcase.
llvm-svn: 176032
Matt Beaumont-Gay [Mon, 25 Feb 2013 18:11:18 +0000 (18:11 +0000)]
'Hexadecimal' has two 'a's and only one 'i'.
llvm-svn: 176031
Bill Schmidt [Mon, 25 Feb 2013 16:44:35 +0000 (16:44 +0000)]
Fix missing relocation for TLS addressing peephole optimization.
Report and fix due to Kai Nacke. Testcase update by me.
llvm-svn: 176029
Daniel Jasper [Mon, 25 Feb 2013 15:59:54 +0000 (15:59 +0000)]
Re-add hack that caused regression.
This fixes llvm.org/PR15350.
Before:
Constructor(int Parameter = 0)
:
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa(
aaaaaaaaaaaaaaaaa),
aaaaaaaaaaaa(
aaaaaaaaaaaaaaaaa) {}
After:
Constructor(int Parameter = 0)
:
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa(
aaaaaaaaaaaaaaaaa),
aaaaaaaaaaaa(
aaaaaaaaaaaaaaaaa) {}
I think the correct solution is to put the VariablePos into
ParenState, not LineState. Added FIXME.
llvm-svn: 176027
Howard Hinnant [Mon, 25 Feb 2013 15:50:36 +0000 (15:50 +0000)]
Alexey Samsonov: #ifdefs out undefined function in static build of libc++ w/o RTTI.
llvm-svn: 176026
Edwin Vane [Mon, 25 Feb 2013 14:49:29 +0000 (14:49 +0000)]
Adding hasDeclaration overload for TemplateSpecializationType
TemplateSpecializationType doesn't quite have getDecl(). Need to go
through TemplateName to get a TemplateDecl.
Added test cases for the hasDeclaration() overload for
TemplateSpecializationType. Also introduced the type matcher
templateSpecializationType() used by the new hasDeclaration() test case.
Updated LibASTMatchersReference.
Reviewers: klimek
llvm-svn: 176025
Edwin Vane [Mon, 25 Feb 2013 14:32:42 +0000 (14:32 +0000)]
Support in hasDeclaration for types with getDecl()
Re-introducing r175532. The has_getDecl metafunction didn't compile with
Visual Studio. This revision uses approaches has_getDecl from a
different angle that isn't a problem for Visual Studio.
Added dedicated tests for the metafunction.
Reviewers: klimek
llvm-svn: 176024
Chandler Carruth [Mon, 25 Feb 2013 14:29:38 +0000 (14:29 +0000)]
Fix spelling noticed by Duncan.
llvm-svn: 176023
Chandler Carruth [Mon, 25 Feb 2013 14:20:21 +0000 (14:20 +0000)]
Fix the root cause of PR15348 by correctly handling alignment 0 on
memory intrinsics in the SDAG builder.
When alignment is zero, the lang ref says that *no* alignment
assumptions can be made. This is the exact opposite of the internal API
contracts of the DAG where alignment 0 indicates that the alignment can
be made to be anything desired.
There is another, more explicit alignment that is better suited for the
role of "no alignment at all": an alignment of 1. Map the intrinsic
alignment to this early so that we don't end up generating aligned DAGs.
It is really terrifying that we've never seen this before, but we
suddenly started generating a large number of alignment 0 memcpys due to
the new code to do memcpy-based copying of POD class members. That patch
contains a bug that rounds bitfield alignments down when they are the
first field. This can in turn produce zero alignments.
This fixes weird crashes I've seen in library users of LLVM on 32-bit
hosts, etc.
llvm-svn: 176022
Edwin Vane [Mon, 25 Feb 2013 14:09:28 +0000 (14:09 +0000)]
Reverting r176019. Corrupt patch.
llvm-svn: 176021
Alexey Samsonov [Mon, 25 Feb 2013 14:06:38 +0000 (14:06 +0000)]
[Sanitizer] one more fix for signed/unsigned mismatch in comparison
llvm-svn: 176020
Edwin Vane [Mon, 25 Feb 2013 13:59:06 +0000 (13:59 +0000)]
Support in hasDeclaration for types with getDecl()
Re-introducing r175532. The has_getDecl metafunction didn't compile with Visual
Studio. This revision approaches has_getDecl from a different angle that isn't
a problem for Visual Studio.
Added dedicated tests for the metafunction.
Reviewers: klimek
llvm-svn: 176019
Daniel Jasper [Mon, 25 Feb 2013 12:39:41 +0000 (12:39 +0000)]
Update generated documentation after recent matcher addition.
llvm-svn: 176018
Daniel Jasper [Mon, 25 Feb 2013 12:02:08 +0000 (12:02 +0000)]
Add matcher for AccessSpecDecls.
Also, add matchers isPrivate(), isProtected() and isPublic(), that
restrict the matching of such AccessSpecDecls and all other Decls.
llvm-svn: 176017
Alexey Samsonov [Mon, 25 Feb 2013 09:00:03 +0000 (09:00 +0000)]
[Sanitizer] fix signed/unsigned compare warnings in sanitizer_stacktrace_test. Build sanitizer_common tests with -Werror as we use predictable just-built Clang for building them
llvm-svn: 176014
Alexey Samsonov [Mon, 25 Feb 2013 08:43:10 +0000 (08:43 +0000)]
[TSan] Add interceptor for malloc_usable_size()
llvm-svn: 176013
Alexey Samsonov [Mon, 25 Feb 2013 08:32:02 +0000 (08:32 +0000)]
[TSan] remove temporary build directory in 'make clean' command
llvm-svn: 176012
Ted Kremenek [Mon, 25 Feb 2013 07:37:18 +0000 (07:37 +0000)]
[analyzer] Relax assumption in FindLastStoreBRVisitor that the thing we are looking for is always a VarRegion.
This was triggering assertion failures when analyzing the LLVM codebase. This
is fallout from r175988.
I've got delta chewing away on a test case, but I wanted the fix to go
in now.
llvm-svn: 176011
Ted Kremenek [Mon, 25 Feb 2013 07:37:13 +0000 (07:37 +0000)]
[analyzer] add the notion of an "interesting" lvalue expression for ExplodedNode pruning.
r175988 modified the ExplodedGraph trimming algorithm to retain all
nodes for "lvalue" expressions. This patch refines that notion to
only "interesting" expressions that would be used for diagnostics.
llvm-svn: 176010
Bill Wendling [Mon, 25 Feb 2013 07:15:16 +0000 (07:15 +0000)]
Add more attributes from the command line to functions.
This is an ongoing process. Any command line option which a back-end cares about
should be added here.
llvm-svn: 176009
Han Ming Ong [Mon, 25 Feb 2013 05:41:42 +0000 (05:41 +0000)]
<rdar://problem/
13282582>
Need available CPU on target device to support CPU reporting.
llvm-svn: 176008
Reed Kotler [Mon, 25 Feb 2013 02:25:47 +0000 (02:25 +0000)]
Make pseudos FEXT_CCRX16_ins and FEXT_CCRXI16_ins into custom emitters.
llvm-svn: 176007
Dmitri Gribenko [Mon, 25 Feb 2013 01:14:45 +0000 (01:14 +0000)]
Documentation: add an idea for a (weird) clang tool
llvm-svn: 176005
David Blaikie [Mon, 25 Feb 2013 01:07:18 +0000 (01:07 +0000)]
DIBuilder: support structs with vtable pointers in the same way as classes
llvm-svn: 176004
David Blaikie [Mon, 25 Feb 2013 01:07:08 +0000 (01:07 +0000)]
Update Clang for a minor DIBuilder breaking change.
llvm-svn: 176003
Reed Kotler [Sun, 24 Feb 2013 23:17:51 +0000 (23:17 +0000)]
Make psuedo FEXT_T8I816_ins into a custom emitter.
llvm-svn: 176002
Aaron Ballman [Sun, 24 Feb 2013 19:08:10 +0000 (19:08 +0000)]
Moving initialization into the initializer list and correcting capitalization, as suggested by Chandler.
llvm-svn: 176001
Dmitri Gribenko [Sun, 24 Feb 2013 19:04:36 +0000 (19:04 +0000)]
Documentation: add an idea for an automatic refactoring
llvm-svn: 176000
Daniel Jasper [Sun, 24 Feb 2013 18:54:32 +0000 (18:54 +0000)]
Allow breaking between a type and name in variable declarations.
This fixes llvm.org/PR14967 and is generall necessary to avoid
situations where the column limit is exceeded. The challenge is
restricting such lines splits, otherwise clang-format suddenly starts
breaking at bad places.
Before:
ReallyLongReturnType<TemplateParam1, TemplateParam2>
ReallyReallyLongFunctionName(
const std::string &SomeParameter,
const SomeType<string,
SomeOtherTemplateParameter> &ReallyReallyLongParameterName,
const SomeType<string,
SomeOtherTemplateParameter> &AnotherLongParameterName) {}
After:
ReallyLongReturnType<TemplateParam1, TemplateParam2>
ReallyReallyLongFunctionName(
const std::string &SomeParameter,
const SomeType<string, SomeOtherTemplateParameter> &
ReallyReallyLongParameterName,
const SomeType<string, SomeOtherTemplateParameter> &
AnotherLongParameterName) {}
llvm-svn: 175999
Bill Schmidt [Sun, 24 Feb 2013 17:34:50 +0000 (17:34 +0000)]
Fix PR14364.
This removes a const_cast hack from PPCRegisterInfo::hasReservedSpillSlot().
The proper place to save the frame index for the CR spill slot is in the
PPCFunctionInfo object, not the PPCRegisterInfo object.
No new test cases, as this just reimplements existing function. Existing
tests such as test/CodeGen/PowerPC/crsave.ll are sufficient.
llvm-svn: 175998
Shankar Easwaran [Sun, 24 Feb 2013 16:02:24 +0000 (16:02 +0000)]
[ELF][Hexagon] add _SDA_BASE_(absolute symbol pointing to start of quickdata)
llvm-svn: 175997
Benjamin Kramer [Sun, 24 Feb 2013 15:51:32 +0000 (15:51 +0000)]
Add the testcase from PR13573, this used to crash.
The error is a bit strange tbh, but better than crashing.
llvm-svn: 175996
Benjamin Kramer [Sun, 24 Feb 2013 15:34:43 +0000 (15:34 +0000)]
CVP: If we have a PHI with an incoming select, try to skip the select.
This is a common pattern with dyn_cast and similar constructs, when the
PHI no longer depends on the select it can often be turned into a simpler
construct or even get hoisted out of the loop.
PR15340.
llvm-svn: 175995
Benjamin Kramer [Sun, 24 Feb 2013 15:34:29 +0000 (15:34 +0000)]
Fix invalid IR in test, missing incoming value for PHI node.
llvm-svn: 175994
Shankar Easwaran [Sun, 24 Feb 2013 15:11:55 +0000 (15:11 +0000)]
[ELF] Change functions to virtual, in ExecutableAtoms
llvm-svn: 175993
Shankar Easwaran [Sun, 24 Feb 2013 14:54:02 +0000 (14:54 +0000)]
[ELF]Change name of runtime files, so that targets can change runtime files that they own
llvm-svn: 175992
Francois Pichet [Sun, 24 Feb 2013 12:34:13 +0000 (12:34 +0000)]
Typo
llvm-svn: 175991
Ted Kremenek [Sun, 24 Feb 2013 07:21:01 +0000 (07:21 +0000)]
[analyzer] tracking stores/constraints now works for ObjC ivars or struct fields.
This required more changes than I originally expected:
- ObjCIvarRegion implements "canPrintPretty" et al
- DereferenceChecker indicates the null pointer source is an ivar
- bugreporter::trackNullOrUndefValue() uses an alternate algorithm
to compute the location region to track by scouring the ExplodedGraph.
This allows us to get the actual MemRegion for variables, ivars,
fields, etc. We only hand construct a VarRegion for C++ references.
- ExplodedGraph no longer drops nodes for expressions that are marked
'lvalue'. This is to facilitate the logic in the previous bullet.
This may lead to a slight increase in size in the ExplodedGraph,
which I have not measured, but it is likely not to be a big deal.
I have validated each of the changed plist output.
Fixes <rdar://problem/
12114812>
llvm-svn: 175988
Ted Kremenek [Sun, 24 Feb 2013 07:20:53 +0000 (07:20 +0000)]
Add "KnownSVal" to represent SVals that cannot be UnknownSVal.
This provides a few sundry cleanups, and allows us to provide
a compile-time check for a case that was a runtime assertion.
llvm-svn: 175987
Ted Kremenek [Sun, 24 Feb 2013 07:20:47 +0000 (07:20 +0000)]
Add regression test for serialized diagnostics for notes without locations.
This meant to be included in r175802.
llvm-svn: 175986
Nadav Rotem [Sun, 24 Feb 2013 07:09:35 +0000 (07:09 +0000)]
Revert r169638 because it broke Mesa llvmpipe tests.
Fix PR15239.
llvm-svn: 175985
Reed Kotler [Sun, 24 Feb 2013 06:16:39 +0000 (06:16 +0000)]
Make psuedo FEXT_T8I816_ins a custom inserter. It should be expanded
as early as possible; which means during instruction selection.
llvm-svn: 175984
Shankar Easwaran [Sun, 24 Feb 2013 03:09:10 +0000 (03:09 +0000)]
[ELF][Hexagon]add typeZeroFillFast
llvm-svn: 175983
Richard Smith [Sun, 24 Feb 2013 01:56:24 +0000 (01:56 +0000)]
PR15338: Don't assert if -fsanitize=bounds sees array indexing on an incomplete
array type.
llvm-svn: 175982
Cameron Zwarich [Sun, 24 Feb 2013 01:26:05 +0000 (01:26 +0000)]
Add a use of an otherwise unused variable to remove a warning in non-Asserts
builds.
llvm-svn: 175981
Cameron Zwarich [Sun, 24 Feb 2013 00:27:29 +0000 (00:27 +0000)]
TwoAddressInstructionPass::tryInstructionTransform() only potentially returns
true when shouldOnlyCommute is false, so we can remove code that checks
otherwise.
llvm-svn: 175980
Cameron Zwarich [Sun, 24 Feb 2013 00:27:26 +0000 (00:27 +0000)]
TwoAddrInstructionPass::tryInstructionTransform() has a case where it calls
itself recursively with a new instruction that has not been finalized, in order
to determine whether to keep the instruction. On 'make check' and test-suite the
only cases where the recursive invocation made any transformations were simple
instruction commutations, so I am restricting the recursive invocation to do
only this.
The other cases wouldn't work correctly when updating LiveIntervals, since the
new instructions don't have slot indices and LiveIntervals hasn't yet been
updated. If the other transformations were actually triggering in any test case
it would be possible to support it with a lot of effort, but since they don't
it's not worth it.
llvm-svn: 175979
Argyrios Kyrtzidis [Sun, 24 Feb 2013 00:05:14 +0000 (00:05 +0000)]
[preprocessor] Use MacroDirective in the preprocessor callbacks to make available the
full information about the macro (e.g if it was imported and where).
llvm-svn: 175978
Argyrios Kyrtzidis [Sun, 24 Feb 2013 00:05:05 +0000 (00:05 +0000)]
Add comments for the MacroDirective & MacroInfo classes.
llvm-svn: 175977
Argyrios Kyrtzidis [Sun, 24 Feb 2013 00:05:01 +0000 (00:05 +0000)]
In Sema::InstantiateStaticDataMemberDefinition, pass the var decl to the consumer
just using ASTConsumer::HandleCXXStaticMemberVarInstantiation(), don't pass it with
ASTConsumer::HandleTopLevelDecl.
ASTConsumer::HandleTopLevelDecl is intended for user-written top-level decls;
a consumer can treat an instantiated static data member however it wants of course.
llvm-svn: 175976
Reed Kotler [Sat, 23 Feb 2013 23:37:03 +0000 (23:37 +0000)]
Add new base instruction def for cmpi, cmp, slt and sltu so that def/uses
proper. Fixed this already a few days ago for slti.
llvm-svn: 175975
Cameron Zwarich [Sat, 23 Feb 2013 23:13:28 +0000 (23:13 +0000)]
TargetInstrInfo::commuteInstruction() doesn't actually return a new instruction
unless it was requested to with an optional parameter that defaults to false, so
we don't need to handle that case in TwoAddressInstructionPass.
llvm-svn: 175974
Daniel Jasper [Sat, 23 Feb 2013 21:01:55 +0000 (21:01 +0000)]
Better formatting of conditional expressions.
In conditional expressions, if the condition is split over multiple
lines, also break before both operands.
This prevents formattings like:
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaa ==
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa ? b : c;
Which are bad, because they suggestion incorrect operator precedence:
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaa ==
(
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa ? b : c);
This lead to the discovery that the expression parser incorrectly
handled conditional operators and that it could also handle semicolons
(which in turn reduced the amount of special casing for for-loops). As a
side-effect, we can now apply the bin-packing configuration to the
sections of for-loops.
llvm-svn: 175973
Michael J. Spencer [Sat, 23 Feb 2013 19:46:18 +0000 (19:46 +0000)]
[ELF][Writer] Add hash table.
llvm-svn: 175972
David Blaikie [Sat, 23 Feb 2013 19:30:31 +0000 (19:30 +0000)]
Reference the canonical commit (r175462) in the value casting release note.
llvm-svn: 175971
David Blaikie [Sat, 23 Feb 2013 19:27:10 +0000 (19:27 +0000)]
Document the value casting API changes.
After changing the way several value inheritance hierarchies (TypeLoc,
CFGelement, ProgramPoint and SVal) handle casting, this documentation describes
how 3rd party code may need to be updated to compile with the new APIs.
As suggested by Sean Silva on cfe-dev.
llvm-svn: 175970
David Blaikie [Sat, 23 Feb 2013 19:20:56 +0000 (19:20 +0000)]
Prefer 'and' over '&' in comments.
Post commit code review feedback from Matt Beaumont-Gay on r174248.
llvm-svn: 175969
Benjamin Kramer [Sat, 23 Feb 2013 16:56:22 +0000 (16:56 +0000)]
Shrink SmallPtrSet. It gets swapped a lot which copies the whole small part.
Testing shows that it's empty in >99% of the cases and I couldn't find a case
where it contained more than 2 elements.
llvm-svn: 175967
Benjamin Kramer [Sat, 23 Feb 2013 13:53:57 +0000 (13:53 +0000)]
Add streamed versions of getQualifiedNameAsString.
Move the cold virtual method getNameForDiagnostic out of line.
llvm-svn: 175966
Renato Golin [Sat, 23 Feb 2013 13:25:41 +0000 (13:25 +0000)]
Add global structure vectorization to docs
llvm-svn: 175965
Renato Golin [Sat, 23 Feb 2013 12:48:30 +0000 (12:48 +0000)]
Some more tests for the global structure vectorizer
llvm-svn: 175964
Enea Zaffanella [Sat, 23 Feb 2013 12:27:19 +0000 (12:27 +0000)]
Remove data member MSAsmStmt::AsmLoc, wrongly hiding AsmStmt::AsmLoc.
llvm-svn: 175963
Benjamin Kramer [Sat, 23 Feb 2013 10:40:58 +0000 (10:40 +0000)]
X86: Disable cmov-memory patterns on subtargets without cmov.
Fixes PR15115.
llvm-svn: 175962