Commit Graph

139837 Commits

Author SHA1 Message Date
Hal Finkel 30e95a8ebb BBVectorize: Use VTTI to compute costs for intrinsics vectorization
For the time being this includes only some dummy test cases. Once the
generic implementation of the intrinsics cost function does something other
than assuming scalarization in all cases, or some target specializes the
interface, some real test cases can be added.

Also, for consistency, I changed the type of IID from unsigned to Intrinsic::ID
in a few other places.

llvm-svn: 171079
2012-12-26 01:36:57 +00:00
Craig Topper caef1c5d86 Remove alignment requirement from VCVTSS2SD in folding tables. Reverting r171049. This instruction doesn't require alignment.
llvm-svn: 171078
2012-12-26 00:35:47 +00:00
Rafael Espindola dfa5f9b511 Use the most recent redecl to decide if it is needed.
This fixes pr14691, which I think is a regression from r168519.

llvm-svn: 171077
2012-12-26 00:13:29 +00:00
Hal Finkel b44f890133 LoopVectorize: Enable vectorization of the fmuladd intrinsic
llvm-svn: 171076
2012-12-25 23:21:29 +00:00
Hal Finkel 2a456112ec BBVectorize: Enable vectorization of the fmuladd intrinsic
llvm-svn: 171075
2012-12-25 22:36:08 +00:00
Richard Smith 940a6d73ce Produce an actual error before attempting to attach notes to it when bailing out
due to FORCE_CLANG_DIAGNOSTICS_CRASH=1. Also add a test for that env var.

llvm-svn: 171074
2012-12-25 21:56:27 +00:00
Hal Finkel 2ebe6d08cd Loosen scheduling restrictions on the PPC dcbt intrinsic
As with the prefetch intrinsic to which it maps, simply have dcbt
marked as reading from and writing to its arguments instead of having
unmodeled side effects. While this might cause unwanted code motion
(because aliasing checks don't really capture cache-line sharing),
it is more important that prefetches in unrolled loops don't block
the scheduler from rearranging the unrolled loop body.

llvm-svn: 171073
2012-12-25 18:51:18 +00:00
Hal Finkel 1b5ff08d43 Expand PPC64 atomic load and store
Use of store or load with the atomic specifier on 64-bit types would
cause instruction-selection failures. As with the 32-bit case, these
can use the default expansion in terms of cmp-and-swap.

llvm-svn: 171072
2012-12-25 17:22:53 +00:00
Evgeniy Stepanov 84f46d9993 [msan] Improve recvfrom interceptor.
llvm-svn: 171071
2012-12-25 16:51:57 +00:00
Evgeniy Stepanov f19c086d1e [msan] Fix handling of vectors of pointers.
VectorType::getInteger() can not be used with them, because pointer size
depends on the target.

llvm-svn: 171070
2012-12-25 16:04:38 +00:00
Evgeniy Stepanov ec8371283b [msan] Fix handling of select with vector condition.
llvm-svn: 171069
2012-12-25 14:56:21 +00:00
Erik Verbruggen 11a2eccc8b Fix for PR12222.
Changed getLocStart() and getLocEnd() to be required for Stmts, and make
getSourceRange() optional. The default implementation for getSourceRange()
is build the range by calling getLocStart() and getLocEnd().

llvm-svn: 171067
2012-12-25 14:51:39 +00:00
Benjamin Kramer a9f265ee98 Harden test so it's not affected by changes to compare lowering.
This only failed on hosts that don't have SSE41.

llvm-svn: 171066
2012-12-25 13:23:23 +00:00
Evgeniy Stepanov a57662d7c4 Add MSan unit tests to check-msan and check-all.
llvm-svn: 171065
2012-12-25 13:10:15 +00:00
Benjamin Kramer 81b5a8fd2e X86: Shave off one shuffle from the pcmpeqq sequence for SSE2 by making use of and commutativity.
llvm-svn: 171064
2012-12-25 13:09:08 +00:00
Benjamin Kramer df4af41b9b X86: Custom lower <2 x i64> eq and ne when SSE41 is not available.
pcmpeqd, pshufd, pshufd, pand is cheaper than unpack + cmpq, sbbq, cmpq, sbbq + pack.
Small speedup on loop-vectorized viterbi (-march=core2).

llvm-svn: 171063
2012-12-25 12:54:19 +00:00
Evgeniy Stepanov f45d92bb22 MemorySanitizer unit tests.
llvm-svn: 171062
2012-12-25 12:39:56 +00:00
Alexey Samsonov 788381b8ac ASan: initialize callbacks from ASan module pass in a separate function for consistency
llvm-svn: 171061
2012-12-25 12:28:20 +00:00
Alexey Samsonov 1e3f7ba8f7 ASan: move stack poisoning logic into FunctionStackPoisoner struct
llvm-svn: 171060
2012-12-25 12:04:36 +00:00
Evgeniy Stepanov 6722716d78 [msan] Fix header comments.
llvm-svn: 171059
2012-12-25 11:53:51 +00:00
Kostya Serebryany eff9b4b9b1 [asan] add a test for right OOB with special large sizes. Fix this test in asan_allocator2. More test tweaking for allocator2
llvm-svn: 171058
2012-12-25 09:40:20 +00:00
Dmitry Vyukov 116518af74 tsan: fix tests to consistently use internal functions (otherwise it leads to false positives)
llvm-svn: 171057
2012-12-25 09:20:58 +00:00
Guy Benyei f0a014bbb7 Add intel_ocl_bicc calling convention as a function attribute to clang. The calling convention is already implemented in LLVM.
llvm-svn: 171056
2012-12-25 08:53:55 +00:00
Kostya Serebryany ad24c58e16 [sanitizer] increase the maximum size class of the fast allocator. This should save quite a bit of memory in tsan/msan (and later in asan). This also puts more stress on the large allocator. Add a couple of checks
llvm-svn: 171055
2012-12-25 07:50:35 +00:00
Dmitry Vyukov 774defc649 tsan: fix the code that searches for heap memory block metadata
(the old code won't work with new allocator)

llvm-svn: 171054
2012-12-25 07:45:40 +00:00
Rafael Espindola 54606d57a2 Add 171048 back but invalidate the cache of all redeclarations when setting
the body of a functions. The problem was that hasBody looks at the entire chain
and causes problems to -fvisibility-inlines-hidden if the cache was not
invalidated.

Original message:

Cache visibility of decls.

This unifies the linkage and visibility caching. I first implemented this when
working on pr13844, but the previous fixes removed the performance advantage of
this one.
This is still a step in the right direction for making linkage and visibility
cheap to use.

llvm-svn: 171053
2012-12-25 07:31:49 +00:00
Kostya Serebryany 181f504d82 [asan] properly define ASAN_LOW_MEMORY. Due to a mistake it has been always true causing redzone and quarantine sizes to be smaller than it was some time before (and out of sync with the documentation). Also make one test less greedy
llvm-svn: 171052
2012-12-25 07:17:17 +00:00
Nick Lewycky d192517cf3 Fix whitespace. No functionality change.
llvm-svn: 171051
2012-12-25 06:13:25 +00:00
NAKAMURA Takumi eb63feb5bd Revert r171048, "Cache visibility of decls."
It broke stage2.

llvm-svn: 171050
2012-12-25 04:47:44 +00:00
Nadav Rotem 00410ae625 VCVTSS2SD requires a strict alignment. Thanks Elena.
llvm-svn: 171049
2012-12-25 03:29:18 +00:00
Rafael Espindola f92a59f0cb Cache visibility of decls.
This unifies the linkage and visibility caching. I first implemented this when
working on pr13844, but the previous fixes removed the performance advantage of
this one.

This is still a step in the right direction for making linkage and visibility
cheap to use.

llvm-svn: 171048
2012-12-25 00:39:58 +00:00
Bob Wilson fe73ac34c5 Rename LLVMContext diagnostic handler types and functions.
These are now generally used for all diagnostics from the backend, not just
for inline assembly, so this drops the "InlineAsm" from the names.  No
functional change.  (I've left aliases for the old names but only for long
enough to let me switch over clang to use the new ones.)

llvm-svn: 171047
2012-12-25 00:07:12 +00:00
NAKAMURA Takumi 04a664e92e [CMake] AddLLVM.cmake: Tweak the corner case that "check-all" doesn't have any tests.
"check-all" can be executed with 0 status, "check-all does nothing, no tools built."

LLVM_EXTERNAL_CLANG_BUILD=OFF LLVM_BUILD_TOOLS=OFF can reproduce this.
Oscar Fuentes reported this. Thank you.

llvm-svn: 171046
2012-12-24 22:43:59 +00:00
Nick Lewycky a5d914196c Simplify RUN lines. No functionality change.
llvm-svn: 171045
2012-12-24 21:45:14 +00:00
Nick Lewycky 521e0d59f3 Quiet gcc's -Wparenthesis warning. No functionality change.
llvm-svn: 171044
2012-12-24 19:58:45 +00:00
Nick Lewycky fb43258080 Fix typo "Makre" -> "Make".
llvm-svn: 171043
2012-12-24 19:55:47 +00:00
Benjamin Kramer 9d46110ff1 Use a std::string rather than a dynamically allocated char* buffer.
This affords us to use std::string's allocation routines and use the destructor
for the memory management. Switching to that also means that we can use
operator==(const std::string&, const char *) to perform the string comparison
rather than resorting to libc functionality (i.e. strcmp).

Patch by Saleem Abdulrasool!

Differential Revision: http://llvm-reviews.chandlerc.com/D230

llvm-svn: 171042
2012-12-24 19:23:30 +00:00
Bob Wilson 4ed23578da Add LLVMContext::emitWarning methods and use them. <rdar://problem/12867368>
When the backend is used from clang, it should produce proper diagnostics
instead of just printing messages to errs(). Other clients may also want to
register their own error handlers with the LLVMContext, and the same handler
should work for warnings in the same way as the existing emitError methods.

llvm-svn: 171041
2012-12-24 18:15:21 +00:00
Dmitri Gribenko 06b84eb414 Fix a typo introduced in r168577: FlAGS -> FLAGS (note the lowercase ell)
Now we really pass -Wcovered-switch-default if the compiler supports it.

llvm-svn: 171040
2012-12-24 17:52:48 +00:00
Daniel Jasper 8fbd96855c Let clang-format format itself.
Apply all formatting changes that clang-format would apply to its own source
code. All choices seem to improve readability (or at least not make it worse).
No functional changes.

llvm-svn: 171039
2012-12-24 16:51:15 +00:00
Daniel Jasper 6d822720f0 Penalize tokens with a lower parenthesis level than the start of the line.
This prevents formattings like this (assuming "parameter" doesn't fit the line):
  bool f = someFunction() && someFunctionWithParam(
      parameter) && someOtherFunction();

Here, "parameter" - the start of line 2 - has a parenthesis level of 2, but
there are subsequent tokens ("&&" and "someOtherFunction") with a lower level.
This is bad for readability as "parameter" hides "someOtherFunction". With this
patch, this changes to:
  bool f = someFunction() &&
           someFunctionWithParam(parameter) &&
           someOtherFunction();

llvm-svn: 171038
2012-12-24 16:43:00 +00:00
Dmitri Gribenko a6c2729ccc AutoRegen.sh: update reference to documentation
llvm-svn: 171037
2012-12-24 15:01:59 +00:00
Kostya Serebryany db3250b8b3 [sanitizer] fix gcc warnings, enable one tests under asan_allocator2
llvm-svn: 171036
2012-12-24 14:53:13 +00:00
Kostya Serebryany fc7de2910c [sanitizer] make LargeMmapAllocator::GetBlockBegin faster by not using a linked list
llvm-svn: 171035
2012-12-24 14:35:14 +00:00
Dmitry Vyukov 19969e5045 tsan: work around FileCheck feature that it does not accept empty input
llvm-svn: 171034
2012-12-24 14:02:34 +00:00
Dmitry Vyukov f6d8021520 tsan: remove static reference to glibc internal _dl_get_tls_static_info()
(this breaks RPM packages)

llvm-svn: 171033
2012-12-24 13:43:53 +00:00
Daniel Jasper 2eda23e78a Align RHS after assignments and return statements.
This changes:
  int Result = a +  // force break
      b;
  return Result +  // force break
      5;

To:
  int Result = a +  // force break
               b;
  return Result +  // force break
         5;

llvm-svn: 171032
2012-12-24 13:43:52 +00:00
Kostya Serebryany ebb23c8cfa [sanitizer] rework the size class map in the sanitizer allocator: make the differences between size classes more uniform.
llvm-svn: 171031
2012-12-24 13:41:07 +00:00
Dmitry Vyukov 73dac64b73 tsan: more defensive file descriptor verification
to prevent assertion failures on code like "write(-1, ...)"

llvm-svn: 171030
2012-12-24 11:43:34 +00:00
NAKAMURA Takumi 1b18db7ea3 llvm/test/CodeGen/X86/fold-vex.ll: Add explicit triple.
llvm-svn: 171029
2012-12-24 11:14:06 +00:00