Kevin Enderby [Mon, 22 Jul 2013 21:25:31 +0000 (21:25 +0000)]
Fix the move to/from accumulator register instructions that use a full 64-bit
absolute address encoded in the instruction.
rdar://8612627 and rdar://
14299221
llvm-svn: 186878
Jakub Staszak [Mon, 22 Jul 2013 21:11:30 +0000 (21:11 +0000)]
Change tabs to spaces.
llvm-svn: 186877
Jakub Staszak [Mon, 22 Jul 2013 21:10:42 +0000 (21:10 +0000)]
Fix typo in test.
llvm-svn: 186876
Anders Carlsson [Mon, 22 Jul 2013 21:08:00 +0000 (21:08 +0000)]
Fix a bug in std::fill_n where memset would end up being called in cases when it shouldn’t.
Reviewed by Howard.
llvm-svn: 186875
Michael Gottesman [Mon, 22 Jul 2013 21:05:47 +0000 (21:05 +0000)]
[stackprotector] Changed isNoopBitcast/sameNoopInput to take TargetLoweringBase instead of TargetLowering.
Both functions only use functionality from TargetLoweringBase.
rdar://
13935163
llvm-svn: 186874
Michael Gottesman [Mon, 22 Jul 2013 21:03:56 +0000 (21:03 +0000)]
Revert "cp11-migrate: Integration with LibFormat"
This reverts commit r186866.
This breaks the build and the original author Guillaume Papin
<guillaume.papin@epitech.eu> asked me to revert so he could look at it more with
revane.
llvm-svn: 186873
Enea Zaffanella [Mon, 22 Jul 2013 20:58:30 +0000 (20:58 +0000)]
Revert commit 186833 (no longer needed after the fix in 186859).
llvm-svn: 186872
Ed Maste [Mon, 22 Jul 2013 20:51:08 +0000 (20:51 +0000)]
Fix i386 FreeBSD build
llvm-svn: 186871
Douglas Gregor [Mon, 22 Jul 2013 20:48:33 +0000 (20:48 +0000)]
Make modules depend on the compiler's own module.map, as a proxy for the compiler itself.
The headers in the compiler's own resource include directory are
system headers, which means we don't stat() them eagerly when loading
a module. Use module.map as a proxy for these headers and the compiler
itself. Fixes <rdar://problem/
13856838>.
llvm-svn: 186870
Craig Topper [Mon, 22 Jul 2013 20:46:37 +0000 (20:46 +0000)]
Recommit r186813: More Intel syntax alias fixes. With the addition of suppressing some of the aliases from being emitted by the asm printer.
llvm-svn: 186869
Michael Gottesman [Mon, 22 Jul 2013 20:44:11 +0000 (20:44 +0000)]
[stackprotector] Refactored ssp prologue creation code into its own helper function.
No functionality change.
rdar://
13935163
llvm-svn: 186868
Manman Ren [Mon, 22 Jul 2013 20:28:53 +0000 (20:28 +0000)]
Debug Info Finder: add processScope to actually handle the Scope.
Instead of just adding the scope to the list, we actually handle the scope.
llvm-svn: 186867
Edwin Vane [Mon, 22 Jul 2013 20:26:29 +0000 (20:26 +0000)]
cp11-migrate: Integration with LibFormat
Adding a feature to optionally reformat code changed by the migrator. Like
LibFormat, can choose between built-in styles (LLVM, Mozilla, Google, Chromium)
or use a YAML-format config file.
Author: Guillaume Papin <guillaume.papin@epitech.eu>
llvm-svn: 186866
Ed Maste [Mon, 22 Jul 2013 20:20:55 +0000 (20:20 +0000)]
elf-core: Run-time reg context selection
Instantiate RegisterContextCore... based on getOS() instead of with
compile-time #ifdef-ery.
The assert()s here are unfortunate, but better than crashing with no
explanation. (This would previously happen for an unsupported
architecture, anyhow.)
We should add an equivalent OS and architecture test to
ProcessElfCore::DoLoadCore() and cleanly report the error to the user.
llvm-svn: 186865
Bill Wendling [Mon, 22 Jul 2013 20:15:41 +0000 (20:15 +0000)]
Use the updated name for the attribute.
llvm-svn: 186864
Bill Wendling [Mon, 22 Jul 2013 20:15:21 +0000 (20:15 +0000)]
Recommit r186217 with testcase fix:
Use the function attributes to pass along the stack protector buffer size.
Now that we have robust function attributes, don't use a command line option to
specify the stack protecto buffer size.
llvm-svn: 186863
Ed Maste [Mon, 22 Jul 2013 20:01:34 +0000 (20:01 +0000)]
elf-core: Remove SetTriple workaround
On FreeBSD I see that Arch, Vendor, OS, and Environment are unchanged
after this call (for x86_64, at least), and core debugging works without
it.
If we need to restore it we should be able to switch on
arch.GetTriple.getOS() instead of a compile-time #ifdef.
Thanks to Daniel Malea for testing on Linux.
llvm-svn: 186862
Akira Hatanaka [Mon, 22 Jul 2013 19:30:38 +0000 (19:30 +0000)]
[mips] Fix MipsAsmParser::parseCCRRegs.
Enable parsing all 32 floating point control registers $0-31 and stop trying to
parse floating point condition code register $fcc0. Also, return ParseFail if
the operand being parsed is not in the expected format.
llvm-svn: 186861
Rafael Espindola [Mon, 22 Jul 2013 19:24:34 +0000 (19:24 +0000)]
Mark ELF's first symbol as format specific so that llvm-nm ignores it.
llvm-svn: 186860
Enea Zaffanella [Mon, 22 Jul 2013 19:10:20 +0000 (19:10 +0000)]
Fixed implementation of C89 6.5.7 p3.
Warning should be emitted only for InitListExpr nodes.
llvm-svn: 186859
Matt Arsenault [Mon, 22 Jul 2013 18:59:58 +0000 (18:59 +0000)]
Fix spelling and grammar
llvm-svn: 186858
Matt Arsenault [Mon, 22 Jul 2013 18:58:53 +0000 (18:58 +0000)]
Fix missing const
llvm-svn: 186857
Fariborz Jahanian [Mon, 22 Jul 2013 18:53:45 +0000 (18:53 +0000)]
Objective-C migration: Use NS_OPTIONS when enumerators
have shift/bitwise operators or are power of 2.
llvm-svn: 186856
Akira Hatanaka [Mon, 22 Jul 2013 18:52:22 +0000 (18:52 +0000)]
[mips] Use ADDu instead of OR to copy general purpose registers. Also, delete
the InstAlias pattern which maps "move" to OR to resolve ambiguity in
MatchTable.
llvm-svn: 186855
Andrew Kaylor [Mon, 22 Jul 2013 18:47:24 +0000 (18:47 +0000)]
Adding example source to support MCJIT/Kaleidoscope blog posts.
llvm-svn: 186854
Shuxin Yang [Mon, 22 Jul 2013 18:40:34 +0000 (18:40 +0000)]
Initialize/Register LTO passes to enable flags like -print-after=<lto-pass>
There already have two "dead" functions, initialize{IPO|IPA}, defined for
similar purpose. I decide not to call these two functions for two reasons:
o. they don't cover all LTO passes (which will soon be separated into IPO
and post-IPO passes)
o. We have not yet figured out the right passes and the ordering for IPO
and post-IPO stages, meaning this change is only for the time being.
Since LTO passes are registered, we are now able to print IR before and
after particular point.
For OSX users:
--------------
"...-Wl,-mllvm -Wl,-print-after=<pass-name>" will print IR after the
specified pass.
For Other UNIX with GNU gold linker:
------------------------------------
"-Wl,-plugin-opt=-print-after=<pass-name>" should work.
(NOTE: no need for "-Wl,-mllvm")
Strip "-Wl," if flags are fed directly to linker instead of clang/clang++.
llvm-svn: 186853
Eric Christopher [Mon, 22 Jul 2013 18:26:18 +0000 (18:26 +0000)]
Fix typo.
llvm-svn: 186852
Eric Christopher [Mon, 22 Jul 2013 18:26:15 +0000 (18:26 +0000)]
Formatting.
llvm-svn: 186851
Argyrios Kyrtzidis [Mon, 22 Jul 2013 18:13:54 +0000 (18:13 +0000)]
[arcmt] Only disable ARC in the second compilation if there were actually ARC errors in the checking phase.
rdar://
14490204
llvm-svn: 186850
Richard Smith [Mon, 22 Jul 2013 18:09:32 +0000 (18:09 +0000)]
Add testcase for PR16134, which no longer crashes with ToT.
llvm-svn: 186849
Richard Smith [Mon, 22 Jul 2013 18:06:23 +0000 (18:06 +0000)]
Update comment to refer to core issue number.
llvm-svn: 186848
Aaron Ballman [Mon, 22 Jul 2013 17:55:04 +0000 (17:55 +0000)]
This test was missing its -verify argument.
llvm-svn: 186847
Jim Grosbach [Mon, 22 Jul 2013 17:45:55 +0000 (17:45 +0000)]
MC: mayAffectControlFlow() handling for variadic instructions.
Variadic MC instructions don't note whether the variable operands
are uses or defs, so mayAffectControlFlow() must conservatively
assume they are defs and return true if the PC is in the operand
list.
rdar://
14488628
llvm-svn: 186846
Nadav Rotem [Mon, 22 Jul 2013 17:10:48 +0000 (17:10 +0000)]
Fix an obvious typo in the loop vectorizer where the cost model uses the wrong variable. The variable BlockCost is ignored.
We don't have tests for the effect of if-conversion loops because it requires a big test (that includes if-converted loops) and it is difficult to find and balance a loop to do the right thing.
llvm-svn: 186845
Justin Holewinski [Mon, 22 Jul 2013 17:04:40 +0000 (17:04 +0000)]
[NVPTX] Remove unused prototypes
llvm-svn: 186844
Rui Ueyama [Mon, 22 Jul 2013 16:56:34 +0000 (16:56 +0000)]
Rename ReaderCOFF members to follow the LLD coding style.
llvm-svn: 186843
Edwin Vane [Mon, 22 Jul 2013 16:36:58 +0000 (16:36 +0000)]
cpp11-migrate: Make a VirtualFilesHelper class for the unit tests.
This change makes it easier for new unit tests to create virtual files.
Author: Guillaume Papin <guillaume.papin@epitech.eu>
llvm-svn: 186842
Daniel Jasper [Mon, 22 Jul 2013 16:22:13 +0000 (16:22 +0000)]
Fix bug in clang-format's vim integration cause by r186789.
After the first operation, the buffer contents has changed and thus all
other operations would be invalid. Executing the operations in reversed
order should fix this.
llvm-svn: 186840
Hans Wennborg [Mon, 22 Jul 2013 16:18:13 +0000 (16:18 +0000)]
Option parsing: allow aliases in groups
Option aliases in option groups were previously disallowed by an assert.
As far as I can tell, there was no technical reason for this, and I would
like to be able to put cl.exe compatible options in their own group for Clang,
so let's change the assert.
llvm-svn: 186838
Sergey Matveev [Mon, 22 Jul 2013 16:14:38 +0000 (16:14 +0000)]
[sanitizer] Change strip_path_prefix flag behavior.
Previously (in tools other than TSan) the entire prefix of the path had to mach
the argument. With this change, only some suffix of the prefix has to match.
This is the same way this flag works in TSan.
llvm-svn: 186837
Samuel Benzaquen [Mon, 22 Jul 2013 16:13:57 +0000 (16:13 +0000)]
Add support for overloaded matchers. ie different matcher function signatures with the same name.
Summary:
Add support for overloaded matchers.
This composes with other features, like supporting polymorphic matchers.
Reviewers: klimek
CC: cfe-commits, revane
Differential Revision: http://llvm-reviews.chandlerc.com/D1188
llvm-svn: 186836
Marshall Clow [Mon, 22 Jul 2013 16:02:19 +0000 (16:02 +0000)]
Make tuple's constructor and std::get<>(tuple) constexpr. Final stage of fixing bug #16599. Thanks to Howard for the review and updates.
llvm-svn: 186834
NAKAMURA Takumi [Mon, 22 Jul 2013 15:59:52 +0000 (15:59 +0000)]
c-index-test.c: Let this C89-compliant since r186817.
warning: initializer for aggregate is not a compile-time constant [-Wc99-extensions]
llvm-svn: 186833
Mihai Popa [Mon, 22 Jul 2013 15:49:36 +0000 (15:49 +0000)]
This adds range checking for "ldr Rn, [pc, #imm]" Thumb
instructions. With this patch:
1. ldr.n is recognized as mnemonic for the short encoding
2. ldr.w is recognized as menmonic for the long encoding
3. ldr will map to either short or long encodings depending on the size of the offset
llvm-svn: 186831
Rafael Espindola [Mon, 22 Jul 2013 15:11:51 +0000 (15:11 +0000)]
Replace archive members in the old position.
This matches gnu archive behavior and since archive member order can change
which member is used, not changing the order on replacement looks like the
right thing to do.
This patch also refactors the logic for which archive member to keep and
whether to move it to a helper function (computeInsertAction). The
nesting in computeNewArchiveMembers was getting a bit confusing.
llvm-svn: 186829
Reid Kleckner [Mon, 22 Jul 2013 13:51:44 +0000 (13:51 +0000)]
[ms-cxxabi] Emit linkonce complete dtors in TUs that need them
Based on Peter Collingbourne's destructor patches.
Prior to this change, clang was considering ?1 to be the complete
destructor and the base destructor, which was wrong. This lead to
crashes when clang tried to emit two LLVM functions with the same name.
In this ABI, TUs with non-inline dtors might not emit a complete
destructor. They are emitted as inline thunks in TUs that need them,
and they always delegate to the base dtors of the complete class and its
virtual bases. This change uses the DeferredDecls machinery to emit
complete dtors as needed.
Currently in clang try body destructors can catch exceptions thrown by
virtual base destructors. In the Microsoft C++ ABI, clang may not have
the destructor definition, in which case clang won't wrap the virtual
virtual base destructor calls in a try-catch. Diagnosing this in user
code is TODO.
Finally, for classes that don't use virtual inheritance, MSVC always
calls the base destructor (?1) directly. This is a useful code size
optimization that avoids emitting lots of extra thunks or aliases.
Implementing it also means our existing tests continue to pass, and is
consistent with MSVC's output.
We can do the same for Itanium by tweaking GetAddrOfCXXDestructor, but
it will require further testing.
Reviewers: rjmccall
CC: cfe-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D1066
llvm-svn: 186828
Reid Kleckner [Mon, 22 Jul 2013 13:07:10 +0000 (13:07 +0000)]
Rename D to GD to match the type, which is GlobalDecl
Now we can save GD.getDecl() in D and shorten some casts.
llvm-svn: 186826
Daniel Jasper [Mon, 22 Jul 2013 12:51:47 +0000 (12:51 +0000)]
Install git-clang-format along with clang-format by default.
llvm-svn: 186825
Reid Kleckner [Mon, 22 Jul 2013 12:47:44 +0000 (12:47 +0000)]
Forward declare OMPClause in Sema.h to avoid an include
llvm-svn: 186824
Rafael Espindola [Mon, 22 Jul 2013 12:41:04 +0000 (12:41 +0000)]
Don't depend on a llvm-ar specific behavior.
GNU ar when not given the a or b modifiers replaces archive members in the
same location of the old ones. I am about to implement that in llvm-ar. For
now, just don't depend on the current llvm-ar behavior on this test.
llvm-svn: 186823
Ed Maste [Mon, 22 Jul 2013 12:39:17 +0000 (12:39 +0000)]
Factor out common string match code for POSIX log enable/disable
llvm-svn: 186822
Sergey Matveev [Mon, 22 Jul 2013 12:38:17 +0000 (12:38 +0000)]
[lsan] Make __lsan_do_leak_check() honor the detect_leaks flag.
Also move detect_leaks to common flags.
llvm-svn: 186821
Justin Holewinski [Mon, 22 Jul 2013 12:18:04 +0000 (12:18 +0000)]
[NVPTX] Use approximate FP ops when unsafe-fp-math is used, and append
.ftz to instructions if the nvptx-f32ftz attribute is set to "true"
llvm-svn: 186820
Sergey Matveev [Mon, 22 Jul 2013 11:18:32 +0000 (11:18 +0000)]
[lsan] Print direct leaks first.
Direct leaks are higher priority, so it makes sense to have them on top.
llvm-svn: 186819
Tim Northover [Mon, 22 Jul 2013 11:02:32 +0000 (11:02 +0000)]
Revert "More Intel syntax alias fixes."
This reverts commit r186813, which broke the bots.
llvm-svn: 186818
Enea Zaffanella [Mon, 22 Jul 2013 10:58:26 +0000 (10:58 +0000)]
Implement the part of C89 6.5.7 p3 requiring a constant initializer list
when initializing aggregate/union types, no matter if static or not.
llvm-svn: 186817
Enea Zaffanella [Mon, 22 Jul 2013 10:54:09 +0000 (10:54 +0000)]
Improve clarity/consistency of a few UsingDecl methods and related helpers.
No functionality change.
In Sema helper functions:
* renamed isTypeName as HasTypenameKeyword
In UsingDecl:
* renamed get/setUsingLocation to get/setUsingLoc
* renamed is/setTypeName as has/setTypename
llvm-svn: 186816
Craig Topper [Mon, 22 Jul 2013 10:07:26 +0000 (10:07 +0000)]
Fix typo. Change %cl to CL in Intel pattern.
llvm-svn: 186815
Craig Topper [Mon, 22 Jul 2013 09:58:07 +0000 (09:58 +0000)]
More Intel syntax alias fixes.
llvm-svn: 186814
Craig Topper [Mon, 22 Jul 2013 09:42:31 +0000 (09:42 +0000)]
More Intel syntax alias fixes.
llvm-svn: 186813
Craig Topper [Mon, 22 Jul 2013 09:22:49 +0000 (09:22 +0000)]
Change %xmm0 to XMM0 in Intel side of asm strings for PBLENDVB.
llvm-svn: 186812
Craig Topper [Mon, 22 Jul 2013 09:18:43 +0000 (09:18 +0000)]
Add Intel variants to aliases for some FP instructions.
llvm-svn: 186811
Tim Northover [Mon, 22 Jul 2013 09:06:12 +0000 (09:06 +0000)]
ARM: remove now unneeded custom Asm converters
After Ulrich's r180677 (thanks!) TableGen is intelligent enough to
handle tied constraints involving complex operands properly, so
virtually all of the ARM custom converters are now unnecessary.
llvm-svn: 186810
Craig Topper [Mon, 22 Jul 2013 07:47:51 +0000 (07:47 +0000)]
Reverse operands for Intel syntax form of 'bt' alias.
llvm-svn: 186809
Nadav Rotem [Mon, 22 Jul 2013 05:19:22 +0000 (05:19 +0000)]
Delete unused helper functions.
llvm-svn: 186808
Robert Wilhelm [Mon, 22 Jul 2013 05:04:01 +0000 (05:04 +0000)]
Remove superfluous cast.
llvm-svn: 186807
Tobias Grosser [Mon, 22 Jul 2013 04:11:00 +0000 (04:11 +0000)]
RegisterPass: Unhide -polly-show and -polly-show-only
llvm-svn: 186806
Tobias Grosser [Mon, 22 Jul 2013 03:50:33 +0000 (03:50 +0000)]
ScopDetect: move "return false" out of INVALID macro.
Contributed-by: Star Tan <tanmx_star@yeah.net>
llvm-svn: 186805
Richard Smith [Mon, 22 Jul 2013 03:31:14 +0000 (03:31 +0000)]
PR16288: A template is only missing a default template argument if it provides
any default template arguments, not if it inherits some.
llvm-svn: 186804
Richard Smith [Mon, 22 Jul 2013 02:56:56 +0000 (02:56 +0000)]
Implement DR257 / fix PR16659:
A constructor for an abstract class does not call constructors for virtual
base classes, so it is not an error if no initializer is present for the
virtual base and the virtual base cannot be default initialized.
Also provide a (disabled by default, for now) warning for the case where a
virtual base class's initializer is ignored in an abstract class's constructor,
and address a defect in DR257 where it was not carried through to C++11's rules
for implicit deletion of special member functions.
Based on a patch by Maurice Bos.
llvm-svn: 186803
NAKAMURA Takumi [Mon, 22 Jul 2013 01:27:42 +0000 (01:27 +0000)]
SemaTemplate.cpp: Prune a stray "\param NumParamLists" in comment. [-Wdocumentation]
llvm-svn: 186802
Michael Gottesman [Mon, 22 Jul 2013 00:52:55 +0000 (00:52 +0000)]
Added missing - in the header of PrologEpilogInserter.h so that editors properly realize it is a c++ header and not a c header.
llvm-svn: 186801
Richard Smith [Sun, 21 Jul 2013 23:13:17 +0000 (23:13 +0000)]
Test updates missed from r186799.
llvm-svn: 186800
Richard Smith [Sun, 21 Jul 2013 23:12:18 +0000 (23:12 +0000)]
Tighten up the set of operator new/operator delete calls we're permitted to
optimize, to follow the permissions granted in N3664. Under those rules, only
calls generated by new-expressions and delete-expressions are permitted to be
optimized, and direct calls to ::operator new and ::operator delete must be
treated as normal calls.
llvm-svn: 186799
Richard Smith [Sun, 21 Jul 2013 23:11:42 +0000 (23:11 +0000)]
Treat nothrow forms of ::operator delete and ::operator delete[] as
deallocation functions.
llvm-svn: 186798
Craig Topper [Sun, 21 Jul 2013 22:20:10 +0000 (22:20 +0000)]
Const-correct some iterators. No functional change.
llvm-svn: 186797
Craig Topper [Sun, 21 Jul 2013 21:56:18 +0000 (21:56 +0000)]
Replace 'unsigned short' with 'uint16_t' in a packed data structure for consistency with other fields and to be explicit about bit count.
llvm-svn: 186796
Craig Topper [Sun, 21 Jul 2013 18:58:40 +0000 (18:58 +0000)]
Revert r186649 because it wasn't unnecessary and add a comment.
llvm-svn: 186795
Robert Wilhelm [Sun, 21 Jul 2013 15:20:44 +0000 (15:20 +0000)]
Convert Sema::MatchTemplateParametersToScopeSpecifier to ArrayRef.
llvm-svn: 186794
Lubos Lunak [Sun, 21 Jul 2013 13:15:58 +0000 (13:15 +0000)]
report unused-value warning also for warn_unused types
llvm-svn: 186793
Rafael Espindola [Sun, 21 Jul 2013 12:58:07 +0000 (12:58 +0000)]
Handle replacement into a position past the original member.
We were incorrectly computing where to insert a member if it was replacing
a previous member that was before the insert point.
llvm-svn: 186792
Rafael Espindola [Sun, 21 Jul 2013 12:42:16 +0000 (12:42 +0000)]
Remove unused fields.
llvm-svn: 186791
Benjamin Kramer [Sun, 21 Jul 2013 11:03:40 +0000 (11:03 +0000)]
mem2reg: Minor STL usage cleanup. No functionality change.
llvm-svn: 186790
Daniel Jasper [Sun, 21 Jul 2013 10:45:33 +0000 (10:45 +0000)]
clang-format: Smarter replacement in the vim integration.
With this fix, only changed regions will be replaced in vim's buffer.
Thereby, marks should mostly be left intact. Furthermore, this is a
better fix for the performance problem in conjunction with
'foldmethod=syntax' (see r186660).
llvm-svn: 186789
Chandler Carruth [Sun, 21 Jul 2013 08:37:58 +0000 (08:37 +0000)]
Make the mem2reg interface use an ArrayRef as it keeps a copy of these
to iterate over.
llvm-svn: 186788
Craig Topper [Sun, 21 Jul 2013 07:28:13 +0000 (07:28 +0000)]
Mark that the _ftol2 function used by windows on x86 to handle fptoui modifies ECX.
llvm-svn: 186787
Nadav Rotem [Sun, 21 Jul 2013 06:12:57 +0000 (06:12 +0000)]
Revert a part of r186420. Don't forbid multiple store chains that merge.
llvm-svn: 186786
Logan Chien [Sun, 21 Jul 2013 03:46:55 +0000 (03:46 +0000)]
Fix exception demo: Add mcjit to link component.
Fix exception demo when we are building the examples with configure/make.
This commit updates the link components in the Makefile.
llvm-svn: 186785
Chandler Carruth [Sun, 21 Jul 2013 01:52:33 +0000 (01:52 +0000)]
Hoist the rest of the logic for promoting single-store allocas into the
helper function. This leaves both trivial cases handled entirely in
helper functions and merely manages the list of allocas to process in
the run method.
The next step will be to handle all of the trivial promotion work prior
to even creating the core class and the subsequent simplifications that
enables.
llvm-svn: 186784
Chandler Carruth [Sun, 21 Jul 2013 01:44:07 +0000 (01:44 +0000)]
Hoist the rest of the logic for fully promoting allocas with all uses in
a single block into the helper routine. This takes advantage of the fact
that we can directly replace uses prior to any store with undef to
simplify matters and unconditionally promote allocas only used within
one block.
I've removed the special handling for the case of no stores existing.
This has no semantic effect but might slow things down. I'll fix that in
a later patch when I refactor this entire thing to be easier to manage
the different cases.
llvm-svn: 186783
Chandler Carruth [Sun, 21 Jul 2013 00:01:34 +0000 (00:01 +0000)]
Remove a method made dead by the prior refactoring.
llvm-svn: 186782
Chandler Carruth [Sat, 20 Jul 2013 23:59:51 +0000 (23:59 +0000)]
Hoist the two trivial promotion routines out of the big class that
handles the general cases.
The hope is to refactor this so that we don't end up building the entire
class for the trivial cases. I also want to lift a lot of the early
pre-processing in the initial segment of run() into a separate routine,
and really none of it needs to happen inside the primary promotion
class.
These routines in particular used none of the actual state in the
promotion class, so they don't really make sense as members.
llvm-svn: 186781
Chandler Carruth [Sat, 20 Jul 2013 23:39:26 +0000 (23:39 +0000)]
Hoist the AllocaInfo struct to the top of the file.
This struct is nicely independent of everything else, and we already
needed a foward declaration here. It's simpler to just define it
immediately.
llvm-svn: 186780
Chandler Carruth [Sat, 20 Jul 2013 23:36:19 +0000 (23:36 +0000)]
Sink a typedef and comparator down to the function that actually uses them.
llvm-svn: 186779
Rafael Espindola [Sat, 20 Jul 2013 23:33:15 +0000 (23:33 +0000)]
Don't crash when llvm.compiler.used becomes empty.
GlobalOpt simplifies llvm.compiler.used by removing any members that are also
in the more strict llvm.used. Handle the special case where llvm.compiler.used
becomes empty.
llvm-svn: 186778
Chandler Carruth [Sat, 20 Jul 2013 23:33:06 +0000 (23:33 +0000)]
Don't allocate the DIBuilder on the heap and remove all the complexity
that ensued from that.
llvm-svn: 186777
Chandler Carruth [Sat, 20 Jul 2013 23:23:47 +0000 (23:23 +0000)]
Rename constructor parameters to follow the common member-shadowing
pattern and conform to the naming conventions.
llvm-svn: 186776
Chandler Carruth [Sat, 20 Jul 2013 23:20:08 +0000 (23:20 +0000)]
Reformat the implementation of mem2reg with clang-format so that my
subsequent changes don't introduce inconsistencies.
llvm-svn: 186775
Andrew Trick [Sat, 20 Jul 2013 23:10:31 +0000 (23:10 +0000)]
Comment: try to clarify loop iteration order.
llvm-svn: 186774
Chandler Carruth [Sat, 20 Jul 2013 23:09:05 +0000 (23:09 +0000)]
Remove a DenseMapInfo specialization for std::pair -- we have one of
those baked into DenseMap now.
llvm-svn: 186773